Taste of Research Scholarships

Project Title:	Constraint Programming based Column Generation for Vehicle Routing

Name of Supervisor:	Michael Maher

Email of Supervisor:	Michael.Maher@nicta.com.au

Name of Joint/Co-Supervisor:	Lanbo Zheng

Email of Joint/Co-Supervisor:	Lanbo.Zheng@nicta.com.au

School:	School of Computer Science and Engineering

For CSE and EET Projects:	NICTA Project

Faculty Research Area (Theme):	Spatial Information Systems and Positioning

School Research Area:	Algorithms

Applicable to other Engineering schools/disciplines:

Abstract:	Research Area: Vehicle Routing, Constraint Programming (CP), Column Generation Vehicle routing is one of the most complicated combinatorial optimisation problems and plays a central role in logistics management. Variations of the problem attracted intensive studies in the literature. The aim of this project is to find practical solutions for the capacitated vehicle routing problem with time windows (CVRPTW) by column generation, where the master problem is modeled as an integer programming (IP) problem and we resort to constraint programming to solve the sub-problem. Work will involve developing algorithms based on the mentioned approaches; implementing and testing the algorithms with benchmark data sets.

Research Environment:	You will be working with experts in constraint programming and vehicle routing and get to know the problem from both academic and industrial sides.

Novelty and Contribution:	Hybrid operation research and constraint programming method has become an active topic as they have orthogonal and complementary strengths in stating and solving combinatorial optimisation problems. The success of applying this technique to solve large-scale vehicle routing problem with practical complications will has impact on both theoretical research and commercial applications.

Expected Outcomes:	The student will gain basic knowledge in integrating constraint programming and mathematical programming for vehicle routing. The student will also get familiar with some competitive optimisation tools.

Reference Material Links:	Handbook of Constraint Programming. Editors: F. Rossi, P. Van Beek and T. Walsh. (Chapter 23) Column Generation. Desaulniers, Guy. (Chapter 3) (accessable from the University’s digital library) Solving VRPTWs with Constraint Programming based Column genetation. L. Rousseau, M. Gendreau, G. Pesant and F. Focacci. http://www.springerlink.com/content/x04870404j142822/

Project Title:	Open Global Constraints

Name of Supervisor:	Michael Maher

Email of Supervisor:	michael.maher@nicta.com.au

Name of Joint/Co-Supervisor:

Email of Joint/Co-Supervisor:

School:	School of Computer Science and Engineering

For CSE and EET Projects:	School Project

Faculty Research Area (Theme):	Intelligent & Autonomous Systems

School Research Area:	Algorithms

Applicable to other Engineering schools/disciplines:	Mechanical & Manufacturing Engineering

Abstract:	Constraint programming employs "global constraints" to encapsulate sophisticated propagation algorithms, but each constraint involves a single fixed sequence of variables. Open constraints constrain a dynamic sequence of variables: they allow the addition of variables during execution. This supports the intertwining of problem construction and problem solving, which provides a way to manage the complexity of a constraint problem. While propagators for some open global constraints have been designed, there are many other global constraints that do not yet have an open propagator. In this project you will design, and perhaps implement, propagators for open global constraints. This mainly involves algorithm design and adaptation.

Research Environment:	You will be working closely with a senior researcher.

Novelty and Contribution:	These will be some of the first few open constraints designed and implemented. Implementations are to be linked to the new G12 optimization platform

Expected Outcomes:	Expected outcomes are a report detailing the design of propagators for one or more open constraints, and hopefully an implementation of one.

Reference Material Links:	For constraint programming, there is a rather outdated description here: http://kti.ms.mff.cuni.cz/~bartak/constraints/index.html Many propagators for (closed) global constraints are described in http://www.andrew.cmu.edu/user/vanhoeve/papers/chapter.pdf The design of propagators for some open global constraints is presented in http://www.cse.unsw.edu.au/~mmaher/pubs/cp/open_cpaior.pdf The G12 constraint programming platform is described in http://www.nicta.com.au/research/projects/constraint_programming_platform

Artificial Intelligence

Project Title:	Applying lessons from Netflix challenge to image search

Name of Supervisor:	Arcot Sowmya

Email of Supervisor:	sowmya@cse.unsw.edu.au

Name of Joint/Co-Supervisor:

Email of Joint/Co-Supervisor:

School:	School of Computer Science and Engineering

For CSE and EET Projects:	School Project

Faculty Research Area (Theme):	Intelligent & Autonomous Systems

School Research Area:	Artificial Intelligence

Applicable to other Engineering schools/disciplines:	Biomedical Engineering Electrical Engineering & Telecommunications

Abstract:	The recently awarded 'The Netflix Prize' of $1m was offered by the DVD rental service Netflix for a new algorithm that would substantially improve the accuracy of predictions about how much someone is going to like a movie based on their existing movie preferences. This has brought about substantial amount of research into predictive modelling on very large realworld datasets. The goal for the summer project will be to study the techniques used by the winning entries and to apply them to the problem of automated image annotation and keyword based image search. This will give students an opportunity to become familiar with some of the techniques used for large scale data mining.

Research Environment:	The research team at CSE contains a senior academic and about 7 PhD students / postdocs. The team has solid experience on machine learning in computer vision, with large applications developed for satellite image analysis, medical imaging, motion tracking and activity recognition. These provide the basis to extend to the proposed project area.

Novelty and Contribution:	Nearest-neighbour based methods have proven to be highly effective for large scale image search. It may be possible to achieve state of the art performance.

Expected Outcomes:	Software implementation and demo / publication

Reference Material Links:	'The million dollar programming prize' - IEEE Spectrum http://www.spectrum.ieee.org/computing/software/the-million-dollar-programming-prize A Makadia, V Pavlovic and S Kumar, 'A New Baseline for Image Annotation', European Conference on Computer Vision, 2008 http://www.cs.rutgers.edu/~vladimir/pub/makadia08eccv.pdf

Project Title:	Multi-level Incremental Knowledge Acquisition for Computer Aided Diagnosis of Lung CT Images

Name of Supervisor:	Arcot Sowmya

Email of Supervisor:	sowmya@cse.unsw.edu.au

Name of Joint/Co-Supervisor:	Paul Compton

Email of Joint/Co-Supervisor:	compton@cse.unsw.edu.au

School:	School of Computer Science and Engineering

For CSE and EET Projects:	School Project

Faculty Research Area (Theme):	Intelligent & Autonomous Systems

School Research Area:	Artificial Intelligence

Applicable to other Engineering schools/disciplines:	Biomedical Engineering Electrical Engineering & Telecommunications

Abstract:	Recent developments in computerised tomography (CT) scanning have revolutionized the detection and assessment of lung diseases. Current CT scanners can acquire 150 images in approximately 12 seconds and visualise these in 2 or 3 dimensions. Along with these developments has come the need for automated techniques for in terpretation and analysis. Pleural disease is the commonest manifestation of asb estos exposure, and is currently receiving high attention from medico-legal quar ters. This project is part of a larger ARC funded to develop a a multi-level self-maintaining Computer Aided Diagnosis (CAD) system that incorporates novel computer vision and knowledge acquisition techniques for quantification and assessment of asbestos-related pleural disease (ARPD). The summer project will investigate a multi-level incremental knowledge-based framework, based on Ripple Down Rules (RDR), that can handle multiple expert input and the segmented feature attributes to acquire diagnostic rules. RDR is a well-established technique for obtaining knowledge from domain experts and has been commercialised in different areas. The central idea in RDR is that experts construct rules to deal with individual cases as they occur in normal practice, allowing a knowledge base to gradually evolve. Computer-aided Diagnosis will be based on a cooperating hierarchy of RDRs (called HRDR), where higher-level RDR layers can override conclusions from lower layers. In the summer project. HRDR will be implemented and trained on patient scans, with the input of medical specialists.

Research Environment:	The summer project will be undertaken within the context of the larger ARC funded project, with 3 university researchers from Computer Science and Eng and Medicine, as well as a team of 5 medical specialists (respiratory physicians and radiologists) from St Vincent's and Liverpool hospitals. The main work will be based at the School of CSE, with occasional meetings with the other project members (possibly offsite).

Novelty and Contribution:	Automated methods for the acquisition of knowledge are at the forefront of technology. RDR based CAD systems are ideal to overcome issues in CT image interpretation, which have confounded earlier attempts. They allow clinical data to be seamlessly integrated with image data as required \u2013 similar to how radiologists qualify their interpretation of an image because of other clinical information. Secondly results from different feature detection modules can be combined easily.

Expected Outcomes:	A tool for immediate and accurate assessment of the extent of pleural disease in patients, that will be evaluated in situ by specialists at St Vincent's and Liverpool Hospitals

Reference Material Links:	contact supervisor and co-supervisor

Project Title:	An automated system to enumerate valid data mining processes

Name of Supervisor:	Ghazi Al-Naymat

Email of Supervisor:	ghazi@cse.unsw.edu.au

Name of Joint/Co-Supervisor:	Boualem Benatallah

Email of Joint/Co-Supervisor:	boualem@cse.unsw.edu.au

School:	School of Computer Science and Engineering

For CSE and EET Projects:	School Project

Faculty Research Area (Theme):	Intelligent & Autonomous Systems

School Research Area:	Artificial Intelligence

Applicable to other Engineering schools/disciplines:

Abstract:	The typical data mining (DM) process involves multiple stages. For example, a process may include transforming (pre-processing) data, applying a data mining technique, and post-processing (visualizing) the fine results. Data mining community has developed large number of techniques that perform these typical stages. Thus, designing a system that suggests to the end-user the suitable data mining process is necessary. This project aims to build ontology, which defines the various techniques that used in the mining and visualization processes, and their properties. The system specifies the characteristic of the input data and of the desired mining result, and uses the ontology to enumerate the DM processes that are appropriate (valid) for producing the desired results. Similarly, the appropriate visualization tools will be enumerated. Since there will be number of options of the valid processes, an efficient ranking technique should be used to help users choose between these options.

Research Environment:	The student will work in a team of researchers. A literature review is required to understand the fundamentals of data mining area. The student has to present the system (prototype).

Novelty and Contribution:	- Building a comprehensive ontology that describes the characteristics of the data mining and visualization techniques. - Presenting a prototype that enumerates the valid and ranked data mining processes to produce the desired results.

Expected Outcomes:	A system to enumerate valid DM and visualization techniques. A technical report that will be publishable as a conference or journal paper.

Reference Material Links:	[1] Abraham Bernstein, Foster Provost, and Shawndra Hill. Towards Intelligent Assistance for a Data Mining Process: An Ontology-based Approach for Cost-sensitive Classi?cation. IEEE Transactions on Knowledge and Data Engineering, 17(4):503–518, April 2005. [2] IBM. Many eyes. http://manyeyes.alphaworks.ibm.com/manyeyes/, June 2009. [3]Han, J. & Kamber, M. Data Mining: Concepts and Techniques. The Morgan Kaufmann, 2001 [4] http://rapid-i.com/ [5] http://www.cs.waikato.ac.nz/ml/weka/

Project Title:	An Empirical Evaluation of AI Model Building Systems

Name of Supervisor:	Will Uther

Email of Supervisor:	willu@cse.unsw.edu.au

Name of Joint/Co-Supervisor:

Email of Joint/Co-Supervisor:

School:	School of Computer Science and Engineering

For CSE and EET Projects:	NICTA Project

Faculty Research Area (Theme):	Intelligent & Autonomous Systems

School Research Area:	Artificial Intelligence

Applicable to other Engineering schools/disciplines:

Abstract:	One of the major challenges in building an intelligent agent is having that agent model its environment. That is, starting from just a history sequence of actions the agent has taken and what it saw when it did so, the agent must built a model that allows it to predict the outcomes of future actions. There have been a number of algorithms published recently in the academic literature for this problem that haven't been directly empirically compared against each other (U-Tree, PhiMDP, tPSRs, ...). A student taking this project will implement at least two of these algorithms and compare their performance on standard test problems. For excellent students, there are novel ways these algorithms could be extended and it might be interesting to explore these too.

Research Environment:	The student would be working in the NICTA Kensington "Making Sense of Data" group. There would be opportunity to interact with the authors of at least one of the algorithms.

Novelty and Contribution:	There are two ways in which this project could contribute the state of knowledge. The first, and most important, is that it would be the first empirical comparison of U-Tree and PhiMDP. If an advanced student was also able to extend either of these algorithms, that would also be a great contribution.

Expected Outcomes:	An implementation and empirical comparison of the U-Tree and PhiMDP algorithms.

Reference Material Links:	The U-Tree algorithm: http://portal.acm.org/citation.cfm?id=922924 The Influence/Variance algorithm: http://www.autonlab.org/autonweb/14610.html PhiMDP: http://arXiv.org/abs/0812.4580 tPSR: http://www.cs.cmu.edu/~ggordon/mrosen-ggordon-thrun.tpsr-icml2004.pdf

Project Title:	Artificial Intelligence in Urban Traffic Management

Name of Supervisor:	Chen Cai

Email of Supervisor:	chen.cai@nicta.com.au

Name of Joint/Co-Supervisor:

Email of Joint/Co-Supervisor:

School:	School of Computer Science and Engineering

For CSE and EET Projects:	NICTA Project

Faculty Research Area (Theme):	Signal Processing & Control

School Research Area:	Artificial Intelligence

Applicable to other Engineering schools/disciplines:	Civil & Environmental Engineering Sciences – Maths, Physics, Chemistry

Abstract:	This project applies artificial intelligence techniques to urban traffic management. Managing traffic in urban road network is a challenging and rewarding task. Traffic congestion is a major contributor of social cost, which reached 9.4 billion dollars in Australia in 2005 and projected to rise to 20.4 billion in 2020. Mitigating congestion in the road network relies primarily on traffic lights. In Sydney, traffic light control system covers an area of 800,000 km2 and manages 3664 signlised junctions. Several percents’ improvement in control means substantial savings in social cost. Conventional control methods using preset signal plans lack potential for further improvements. Adaptive systems that make decisions at real-time become the frontier of development, and are being insensitively investigated at NICTA, Australia’s national information and communication centre of excellence. Artificial intelligence techniques are key to establish real-time data processing and dynamic decision-making for adaptive systems. The focus of the project is to apply reinforcement learning to traffic signal control.

Research Environment:	Selected students will work with researchers at NICTA, and participate in the distinguished Smart Transport and Road (STaR) project. Project work includes literature review, group discussion, algorithm development, computer programming and technical presentation. It aims to motivate students to combine theoretical and practical interest in research.

Novelty and Contribution:	Managing traffic lights in road network is a complex problem. Finding its solution could easily become computationally prohibitive because of high dimensionality. Using artificial intelligence techniques we may significantly reduce computational demand, while provide good approximation to the original problem. The project will be part of the ongoing research that aims to upgrade systems operating in the fields.

Expected Outcomes:	The outcome of research is expect to show that adaptive control systems using AI techniques bring improvement in performance and are practical for real-time operation.

Reference Material Links:	Cai, C., Wong, C.K., Heydecker, B.G. (2009) Adaptive traffic signal control using approximate dynamic programming, Transportation Research Part C: Special Issue on AI in Transport Analysis (In Press)

Project Title:	Asbestos-Related Pleural Disease Feature Detection in Lung CT Images

Name of Supervisor:	Arcot Sowmya

Email of Supervisor:	sowmya@cse.unsw.edu.au

Name of Joint/Co-Supervisor:

Email of Joint/Co-Supervisor:

School:	School of Computer Science and Engineering

For CSE and EET Projects:	School Project

Faculty Research Area (Theme):	Intelligent & Autonomous Systems

School Research Area:	Artificial Intelligence

Applicable to other Engineering schools/disciplines:	Biomedical Engineering Electrical Engineering & Telecommunications

Abstract:	Recent developments in computerised tomography (CT) scanning have revolutionized the detection and assessment of lung diseases. Current CT scanners can acquire 150 images in approximately 12 seconds and visualise these in 2 or 3 dimensions. Along with these developments has come the need for automated techniques for interpretation and analysis. Pleural disease is the commonest manifestation of asbestos exposure, and is currently receiving high attention from medico-legal quarters. This project is part of a larger ARC funded to develop a a multi-level self-maintaining Computer Aided Diagnosis (CAD) system that incorporates novel computer vision and knowledge acquisition techniques for quantification and assessment of asbestos-related pleural disease (ARPD). The summer project will investigate automatic feature extraction and quantification techniques for ARPD features, and compare results to current “best practice” techniques for assessment of the clinical consequences of benign ARPD. The goal is to build an automated technique to segment pleural plaques and diffuse pleural thickening in volumetric scans using 3D image analysis and features, and expert input from the medical specialists.

Research Environment:	The summer project will be undertaken within the context of the larger ARC funded project, with 3 university researchers from Computer Science and Eng and Medicine, as well as a team of 5 medical specialists (respiratory physicians and radiologists) from St Vincent's and Liverpool hospitals. The main work will be based at the School of CSE, with occasional meetings with the other project members (possibly offsite).

Novelty and Contribution:	ARPD is likely to become one of the commonest asbestos-related disorders compensated in the next 15 years. It is currently the second commonest abnormality compensated in NSW, representing 30% of all cases compensated in 1994-2005 and a considerable financial burden to industry and the community. This project will use unique clinical data collected from a study conducted within St Vincent’s Hospital that provides scientifically useful data which will also be extremely relevant for clinical and compensation purposes.

Expected Outcomes:	The study will develop for the first time sophisticated, easy-to-use software for detection of DPT and pleural plaques and allow differentiation from asbestos-related malignancy. This is of considerable value for the affected individual and the community.

Reference Material Links:	contact supervisor

Project Title:	Computational Intelligence and Games

Name of Supervisor:	Alan Blair

Email of Supervisor:	blair@cse.unsw.edu.au

Name of Joint/Co-Supervisor:	Malcolm Ryan

Email of Joint/Co-Supervisor:	malcolmr@cse.unsw.edu.au

School:	School of Computer Science and Engineering

For CSE and EET Projects:	School Project

Faculty Research Area (Theme):	Intelligent & Autonomous Systems

School Research Area:	Artificial Intelligence

Applicable to other Engineering schools/disciplines:

Abstract:	Games provide an ideal environment for the study and application of computational intelligence. The aim of this project would be to apply techniques such as neural networks, reinforcement learning or evolutionary computation to a board game, multi-agent game, simulated physical game, or other type of game.

Research Environment:	Working with a small team of researchers at CSE.

Novelty and Contribution:

Expected Outcomes:	The research could potentially lead to publication at a venue such as ICML or IEEE-CIG, and possible entry into one of the associated competitions (e.g. Ms Pac-Man, Infinite Mario, Robotic Car Racing, Simulated Car Racing, etc.)

Reference Material Links:	http://cigames.org, http://www.rl-competition.org

Project Title:	Computational methods to manipulate elections

Name of Supervisor:	Toby Walsh

Email of Supervisor:	tw@cse.unsw.EDU.AU

Name of Joint/Co-Supervisor:	Michael Maher

Email of Joint/Co-Supervisor:	michael.maher@nicta.com.au

School:	School of Computer Science and Engineering

For CSE and EET Projects:	NICTA Project

Faculty Research Area (Theme):	Intelligent & Autonomous Systems

School Research Area:	Artificial Intelligence

Applicable to other Engineering schools/disciplines:

Abstract:	Voting has recently become a topic of study for computer science. For example, suppose I have a group of agents trying to make a collective decision. How do I design a mechanism that encourages them to report their true preferences and not try to manipulate the result? As a second example, suppose I am sitting on a hiring committee. How do I work out which of the other committee members to buy off so my friend is hired? These are all computational quesitons. The aim of this project is to implement and refine algorithms to answer such quesitons. This project will especially suit a student with a liking for discrete mathematics.

Research Environment:	Working with a small team of researchers at NICTA.

Novelty and Contribution:	Many of the algorithms to be explored have never been implemented. The goal will be to write up the results in a research publication.

Expected Outcomes:	Implementation of the new algorithms, research report.

Reference Material Links:	http://www.cse.unsw.edu.au/~tw/waimag07.pdf

Project Title:	Fast Reinforcement Learning from Experience

Name of Supervisor:	Bernhard Hengst

Email of Supervisor:	bernhardh@cse.unsw.du.au

Name of Joint/Co-Supervisor:	Claude Sammut

Email of Joint/Co-Supervisor:	claude@cse.unsw.edu.au

School:	School of Computer Science and Engineering

For CSE and EET Projects:	School Project

Faculty Research Area (Theme):	Intelligent & Autonomous Systems

School Research Area:	Artificial Intelligence

Applicable to other Engineering schools/disciplines:	Sciences – Maths, Physics, Chemistry

Abstract:	Learning from experience is important for autonomous systems as hand-coding behaviours can be difficult and tedious. Biological systems seem to learn very quickly from only a handful of trials. Recent research suggests that we can speed up reinforcement learning by making only weak assumptions about the nature of the system’s dynamics. The aim of this project is to speed up the learning of system dynamics and optimal control policies for unknown non-linear systems. A standard test example is pole-balancing.

Research Environment:	The successful applicant will be working under the supervision of staff with many years of research and industrial experience. The research laboratory is located on the 3rd floor of the CSE building and provides the opportunity to interact with other students working on several related robotics projects.

Novelty and Contribution:	This project aims to develop algorithms in reinforcement learning that can surpass previous attempts at pole-balancing in the speed of learning. The techniques are expected to be transferable, for example, to improve bipedal locomotion for the Nao soccer playing robots, rescue robot articulation in difficult terrains, and for learning smooth grasping and manipulation skills for our humanoid torso robot.

Expected Outcomes:	Machine learning algorithms that are able to learn dynamic skills such as pole-balancing in a minimum of trials.

Reference Material Links:	Introduction to Reinforcement Learning: http://www.cs.ualberta.ca/%7Esutton/book/ebook/the-book.html Probabilistic Inference for Fast Learning and Control, Carl Edward Rasmussen and Marc Peter Deisenroth

Project Title:	Intelligent digital matting

Name of Supervisor:	Arcot Sowmya

Email of Supervisor:	sowmya@cse.unsw.edu.au

Name of Joint/Co-Supervisor:

Email of Joint/Co-Supervisor:

School:	School of Computer Science and Engineering

For CSE and EET Projects:	School Project

Faculty Research Area (Theme):	Intelligent & Autonomous Systems

School Research Area:	Artificial Intelligence

Applicable to other Engineering schools/disciplines:	Biomedical Engineering Electrical Engineering & Telecommunications

Abstract:	Digital matting is the process of extracting a foreground object from an image or video together with an 'alpha channel' or opacity estimate for each pixel covered by the object. This operation enables to seamlessly merge the extracted object with a new background and plays a very important role in many image and video editing applications. The goal of this project will be to develop a system which allows realistic composition of objects from multiple video sequences with minimal user interaction. It will offer students an opportunity to become familiar with advanced video editing techniques.

Research Environment:	The research team at CSE contains a senior academic and about 7 PhD students / postdocs. The team has solid experience on machine learning in computer vision, with large applications developed for satellite image analysis, medical imaging, motion tracking and activity recognition. These provide the basis to extend to the proposed project area.

Novelty and Contribution:	The proposed techniques will be based on novel statistical machine learning.

Expected Outcomes:	Software implementation and demo / publication

Reference Material Links:	J Wang and M Cohen, 'Image and Video Matting: A Survey', Foundations and Trends in Computer Graphics and Vision Vol 3. http://w3.impa.br/~lvelho/ip08/reading/video-matting.pdf

Project Title:	Natural Language Conversation with Agents in an Interactive Virtual World

Name of Supervisor:	Claude Sammut

Email of Supervisor:	claude@cse.unsw.edu.au

Name of Joint/Co-Supervisor:	Maurice Pagnucco

Email of Joint/Co-Supervisor:	morri@cse.unsw.edu.au

School:	School of Computer Science and Engineering

For CSE and EET Projects:	School Project

Faculty Research Area (Theme):	Immersive Systems and Virtual Reality

School Research Area:	Artificial Intelligence

Applicable to other Engineering schools/disciplines:

Abstract:	iCinema’s AVIE is a 360° cylindrical screen with stereo projection that creates a very realistic immersive environment. The aim of this project is to create virtual characters for Avie that a user can interact with using speech and natural language. The project will build on existing software systems, developed in CSE, for creating embodied conversational agents. The scope of the project can vary, depending on the scholar's interests, on aspects of 3D graphics for embodied agents or artificial intelligence, specifically natural language understanding for conversational agents.

Research Environment:	The scholar will work with members of the ARC Centre of Excellence for Autonomous Systems. The software environment includes languages for scripting conversational agents, developed in house. The Scholar will also work in the AVIE installation, linking the conversational agent with the AVIE graphics software.

Novelty and Contribution:	There are two main challenges in this project. The first is to move from the relatively simple 2D cartoon characters we currently use for the conversational agent to 3D characters to integrate smoothly into the VIA environment. The second challenge is to create realistic conversational scripts for characters that act as information sources for AVIE applications such as a tour guide for historical sites.

Expected Outcomes:	Two outcomes are expected: a graphical character that embodies a conversational agent in AVIE and a demonstration for a particular application within AVIE.

Reference Material Links:	Sammut, C. (2001). Managing Context in a Conversational Agent. Linkoping Electronic Articles in Computer and Information Science 6(27). http://www.ida.liu.se/ext/epa/ej/etai/2001/011/01011-etaisep.ps

Project Title:	Natural Language Interface for a Smart Room

Name of Supervisor:	Claude Sammut

Email of Supervisor:	claude@cse.unsw.edu.au

Name of Joint/Co-Supervisor:	Malcolm Ryan

Email of Joint/Co-Supervisor:	malcolmr@cse.unsw.edu.au

School:	School of Computer Science and Engineering

For CSE and EET Projects:	School Project

Faculty Research Area (Theme):	Immersive Systems and Virtual Reality

School Research Area:	Artificial Intelligence

Applicable to other Engineering schools/disciplines:

Abstract:	The "smart room" is a meeting room that has been fitted with various sensors, including cameras and microphones for detecting signals from the human occupants. Devices, such as lights, TVs, radios, DVD players, etc, are under computer control. A speech recognition and natural language understanding system is capable of conversing with the occupants of the room, accepting commands to activate devices or responding to questions about the state of the room. The conversational agent can also use a calendar to schedule events and can access the internet to retrieve information about the news, weather, TV programs, etc. The aim of this project is to extend the user interface to include more multi-modal interaction (i.e. combination of speech and gesture - see related project on vision interaction in the smart room) and to control more complex devices. For example, if the room were a kitchen, controlling cookers, etc. The room may also include a mobile robot.

Research Environment:	The scholar will work with members of the ARC Centre of Excellence for Autonomous Systems. The 'smart room' is on level 4 of the CSE building an contains all the devices mentioned above. The scholar will develop software using a purpose built programming language devised by us and middleware, also developed within CSE.

Novelty and Contribution:	The novel contributions will be in developing multi-modal user interfaces and in developing systems for controlling complex devices such as mobile robots interacting with the room.

Expected Outcomes:	It is expected that the scholar will develop a demonstration system running in the smart room.

Reference Material Links:	Sammut, C. (2001). Managing Context in a Conversational Agent. Linkoping Electronic Articles in Computer and Information Science. 6(27). http://www.ida.liu.se/ext/epa/ej/etai/2001/011/01011-etaisep.ps

Project Title:	Playing Simulated Robotic Soccer

Name of Supervisor:	Bernhard Hengst

Email of Supervisor:	bernhardh@cse.unsw.du.au

Name of Joint/Co-Supervisor:	Claude Sammut

Email of Joint/Co-Supervisor:	claude@cse.unsw.edu.au

School:	School of Computer Science and Engineering

For CSE and EET Projects:	School Project

Faculty Research Area (Theme):	Intelligent & Autonomous Systems

School Research Area:	Artificial Intelligence

Applicable to other Engineering schools/disciplines:

Abstract:	The UNSW has participated at Robocup international autonomous robotic soccer competitions since 1999. Successful teams develop behaviour strategies that emphasise teamwork amongst the robots. Strategies can be tried in simulation, significantly speeding up development time. The aim of this project is to use a team of virtual Naos and develop competitive soccer playing behaviour.

Research Environment:	The successful applicant will be working under the supervision of staff with many years of research and industrial experience. The research laboratory is located on the 3rd floor of the CSE building and houses students working on several related robotics projects. Students may have the opportunity to join the 2010 UNSW Robocup soccer team or participate in Robotstadium, an on-line programming contest based on a simulation of the RoboCup SPL (Standard Platform League). Last year everyone with programming skills could participate in Robotstadium (for free) and win 1000.- Swiss Francs and a Webots software package.

Novelty and Contribution:	This project is designed to give researchers an appreciation for the many challenges in robotics and artificial intelligence. The project has a second aim and that is to develop a competitive autonomous agent game to comprise part of the curriculum for COMP3411 (Introduction to Artificial Intelligence).

Expected Outcomes:	The aim of this project is to program a team of virtual Nao robots to play soccer against each other.

Reference Material Links:	www.robotstadium.org http://www.cyberbotics.com/order.html

Project Title:	Segmentation of Anatomical Landmarks in Lung CT images for Computer Aided Diagnosis

Name of Supervisor:	Arcot Sowmya

Email of Supervisor:	sowmya@cse.unsw.edu.au

Name of Joint/Co-Supervisor:

Email of Joint/Co-Supervisor:

School:	School of Computer Science and Engineering

For CSE and EET Projects:	School Project

Faculty Research Area (Theme):	Intelligent & Autonomous Systems

School Research Area:	Artificial Intelligence

Applicable to other Engineering schools/disciplines:	Biomedical Engineering Electrical Engineering & Telecommunications

Abstract:	Recent developments in computerised tomography (CT) scanning have revolutionized the detection and assessment of lung diseases. Current CT scanners can acquire 150 images in approximately 12 seconds and visualise these in 2 or 3 dimensions. Along with these developments has come the need for automated techniques for interpretation and analysis. Pleural disease is the commonest manifestation of asbestos exposure, and is currently receiving high attention from medico-legal quarters. This project is part of a larger ARC funded to develop a a multi-level self-maintaining Computer Aided Diagnosis (CAD) system that incorporates novel computer vision and knowledge acquisition techniques for quantification and assessment of asbestos-related pleural disease (ARPD). The summer project will investigate novel tissue segmentation techniques incorporating anatomy and diagnostic knowledge that will detect disease features of interest in high resolution CT images of the lung.

Research Environment:	The summer project will be undertaken within the context of the larger ARC funded project, with 3 university researchers from Computer Science and Eng and Medicine, as well as a team of 5 medical specialists (respiratory physicians and radiologists) from St Vincent's and Liverpool hospitals. The main work will be based at the School of CSE, with occasional meetings with the other project members (possibly offsite).

Novelty and Contribution:	Robust detection and segmentation of anatomical landmarks such as ribs, diaphragm and mediastinum are essential for both segmenting disease patterns and image registration between scans in the longitudinal study. While many reported techniques for anatomical segmentation may work well in normal cases, the disease-affected anatomy poses many challenges.

Expected Outcomes:	A tool for immediate and accurate assessment of anatomical landmarks of interest, that will be evaluated in situ by specialists at St Vincent’s and Liverpool Hospitals

Reference Material Links:	contact supervisor

Project Title:	Using GPUs to solve challenging optimization problems

Name of Supervisor:	Toby Walsh

Email of Supervisor:	tw@cse.unsw.EDU.AU

Name of Joint/Co-Supervisor:	Michael Maher

Email of Joint/Co-Supervisor:	michael.maher@nicta.com.au

School:	School of Computer Science and Engineering

For CSE and EET Projects:	NICTA Project

Faculty Research Area (Theme):	Intelligent & Autonomous Systems

School Research Area:	Artificial Intelligence

Applicable to other Engineering schools/disciplines:

Abstract:	GPUs in games machines like the Sony PlayStation and in graphics cards have evolved to the point where non-graphics related computation can be performed on them and may significantly outperform conventional CPUs in some cases. However, most algorithms need to be adapted to take advantage of this modern hardware. The student will work in a small team to develop and test optimization methods to work on a GPU. If time permits, these methods can be used to solve some challenging (even open) optimization problems.

Research Environment:	Working with a small team of researchers at NICTA

Novelty and Contribution:	GPUs have not previously been used in this domain. In addition to practical benefits, new theoretical results may be obtained as algorithms are adapted to take advantage of the parallelism offered by GPUs.

Expected Outcomes:	Implementation of the new algorithms within an existing constraint solver (e.g. the G12 finite domain solver, Minion, others).

Reference Material Links:	http://gpgpu.org/

Project Title:	Using Machine Learning to Model Connections between Political Competition and International Conflict

Name of Supervisor:	Arcot Sowmya

Email of Supervisor:	sowmya@cse.unsw.edu.au

Name of Joint/Co-Supervisor:	Ben Goldfield

Email of Joint/Co-Supervisor:	b.goldsmith@usyd.edu.au

School:	School of Computer Science and Engineering

For CSE and EET Projects:	School Project

Faculty Research Area (Theme):	Intelligent & Autonomous Systems

School Research Area:	Artificial Intelligence

Applicable to other Engineering schools/disciplines:	Biomedical Engineering Electrical Engineering & Telecommunications

Abstract:	The study of connections between domestic politics and international relations has been very active and productive in political science over the past two decades. The most prominent example of this is the “democratic peace” literature, exploring a widely acknowledged empirical finding, that historically democracies have been very unlikely to go to war with each other, although in general they may be no less war-prone than other kinds of state. But explaining why this is the case continues to prove challenging for theorists and empirical analysts. We propose to build on previous work (Goldsmith, Cai, & Sowmya 2008; Goldsmith, Chalup, & Quinlan 2008) to explore a particularly promising hypothesis about the role of institutions for political competition in the democratic peace. Machine learning techniques such as artificial neural networks (ANNs) and support vector machines (SVMs) are especially appropriate for the proposed work because they can capture the contingent and non-linear dynamics of complex political processes much better than standard econometric techniques almost universally employed in the literature. This project can stand alone and lead to a conference paper and/or journal article, but is also designed to complement work proposed in a larger project currently under review as an ARC Discovery Project.

Research Environment:	The summer project will be undertaken within the context of ongoing investigation into these dynamics, which is a collaborative project between the main supervisor, Prof. Arcot Sowmya, and Dr. Benjamin Goldsmith, of the Department of Government & International Relations, University of Sydney.

Novelty and Contribution:	There are very few applications of machine learning to quantitative analysis in political science, but those which have been done include some of the leading empirical researchers and methodologists in the field. Sowmya and Goldsmith have each contributed to this area in the papers cited above. The particular focus on political competition stems from strong findings by Goldsmith (2007) that this aspect of political systems is most strongly connected to behavior closely linked to warfare, especially military spending.

Expected Outcomes:	We will produce a joint conference paper , to be presented at a leading political science and/or computer science/engineering conference. This paper may then be revised for submission to leading journal, either in political science or computer science / engineering.

Reference Material Links:	Goldsmith, B.E. 2007. “Defense Effort and Institutional Theories of Democratic Peace and Victory: Why try harder?” Security Studies 16, 2: 189-222. Goldsmith, B.E., X. Cai & A. Sowmya. 2008. ‘Is International Trade Associated with Peace or War? Some new measures and methods,’ presented at Meeting of American Political Science Assoc., Boston, 28-31 August. Goldsmith, B.E., S.K. Chalup, and M.J. Quinlan. 2008. “Regime Type and International Conflict: Towards a general model,” Journal of Peace Research 45, 6: 743-763. Others available on request

Project Title:	Vision-based hazard detection for the visually impaired

Name of Supervisor:	Arcot Sowmya

Email of Supervisor:	sowmya@cse.unsw.edu.au

Name of Joint/Co-Supervisor:

Email of Joint/Co-Supervisor:

School:	School of Computer Science and Engineering

For CSE and EET Projects:	School Project

Faculty Research Area (Theme):	Intelligent & Autonomous Systems

School Research Area:	Artificial Intelligence

Applicable to other Engineering schools/disciplines:	Biomedical Engineering Electrical Engineering & Telecommunications

Abstract:	Persons with visual impairment usually rely on their hearing system to sense the environment, which also raises the issue of vision-audio-based interactive interfaces. Research in computer vision has make progress on 3-D reconstruction, moving object tracking and activity recognition. These techniques may be used to create intelligent systems to understand the environment and provide aids for hazard prediction, alarming and avoidance. The summer project is part of a larger endeavour to develop a hazard detection system that can actively sense the environment and learn to predict and classify any hazards to the user. It would also provide hazard avoidance facilities as spatial mapping and intelligent navigation, based on computer vision techniques. Cameras are used to capture the scene and construct a 3-D virtual world, in which the motion of objects are tracked, activities are analysed and recognised. The environment information will then be fused and analysed to understand the hazard impact of the environment. The results will be mapped to appropriate signals, such as sounds or vibration, to inform and warn the user.

Research Environment:	The research team at CSE contains a senior academic and about 7 PhD students / postdocs. The team has already developed many motion tracking and activity recognition systems, including for the iCinema AVIE environment as well as face tracking using webcams. All necessary infrastructure is available.

Novelty and Contribution:	Standard approaches for hazard detection use ultrasound sensors and visual cameras to detect objects near the user, which only provide unstructured information on potential barriers. However, without understanding the structure of the environment and activities within it, a hazard detection system is weak. Furthermore, current hazard detection systems concentrate more on the environment, and less on the user. This is unsatisfactory since the degree of attention and response of the user to any potential hazard is very important in analysing the severity of hazards. Developing intelligent hazard detection systems for the visually impaired, that can understand the environment and the user response, is a great challenge.

Expected Outcomes:	A real-time tracking system that converts a spatial map into sound signals, to serve as navigational aid to the visually impaired

Reference Material Links:	contact supervisor

Project Title:	Vision-based head activity and face expression analysis for machine control

Name of Supervisor:	Arcot Sowmya

Email of Supervisor:	sowmya@cse.unsw.edu.au

Name of Joint/Co-Supervisor:

Email of Joint/Co-Supervisor:

School:	School of Computer Science and Engineering

For CSE and EET Projects:	School Project

Faculty Research Area (Theme):	Intelligent & Autonomous Systems

School Research Area:	Artificial Intelligence

Applicable to other Engineering schools/disciplines:	Biomedical Engineering Electrical Engineering & Telecommunications

Abstract:	Within human-machine interaction (HCI) research, great effort is taken to create good user interfaces by directly employing the natural communication and manipulation skills of humans. Adopting vision in HCI will allow the deployment of a wide range of techniques in more sophisticated and restricted scenarios. Such systems are especially important for machine control by people with restricted limb movements. Future HCI interfaces will also need a good understanding of the person’s behaviour so that machines can learn from it, react accordingly, and reproduce this behaviour afterwards. The development of such systems involves addressing of challenging research problems including effective input and output techniques, interaction styles and evaluation methods. In the input domain, the computer vision approach offers the capturing and interpretation of the motion of head, eye gaze, face, hand, arms or even the whole body. The summer project is part of a larger project whose aim is to develop an intelligent vision-based HCI system for humans with restricted limb and body movements, that performs accurate and flexible machine control. It utilises cameras to capture the head pose and movement, as well as face appearance, which are analysed, interpreted and learned into knowledge. The knowledge will then be utilised to control machines.

Research Environment:	The research team at CSE contains a senior academic and about 7 PhD students / postdocs. The team has already developed a learnable level set-based active contour model for image segmentation and recognition and dynamic models for video tracking. These provide a working platform to extend and apply machine learning and active shape and/or appearance models to tackle the design, development and implementation of intelligent HCI for machine control.

Novelty and Contribution:	Face video analysis is often performed through the study of specific face features, such as eyes, eyebrows and mouth, to extract the most significant information regarding expression and speech. Most research on HCI treat head pose and facial expression separately, which prevents the system from using coupled information of head pose and face expression to improve the accuracy in terms of tracking and knowledge presentation. Even when both are used, head information is usually utilised only to predict the pose of the head rather than to study its activity, where the latter provides richer information for HCI. Fusing information of head activity and face expression for machine control is a big challenge.

Expected Outcomes:	Algorithms for one or more of intelligent head tracking, head activity recognition, facial expression recognition

Reference Material Links:	contact supervisor

Autonomous Systems & Sensing Technologies

Project Title:	A multi-touch SmartBoard project

Name of Supervisor:	Arcot Sowmya

Email of Supervisor:	sowmya@cse.unsw.edu.au

Name of Joint/Co-Supervisor:

Email of Joint/Co-Supervisor:

School:	School of Computer Science and Engineering

For CSE and EET Projects:	School Project

Faculty Research Area (Theme):	Immersive Systems and Virtual Reality

School Research Area:	Autonomous Systems & Sensing Technologies

Applicable to other Engineering schools/disciplines:	Biomedical Engineering Electrical Engineering & Telecommunications

Abstract:	In this project, you will implement a multi touch smart board that uses cameras and computer vision techniques to implement multi-touch interaction on a standard whiteboard. The system will combine Wii-mote tracking with webcam techniques to provide a multi-touch interface that can also recognise hand gestures for interaction.

Research Environment:	The research team at CSE contains a senior academic and about 7 PhD students / postdocs. The team has already developed a number of systems for image segmentation and recognition and for video tracking. These provide a working platform to extend and apply th etechnqius to the project.

Novelty and Contribution:	This research explores a unique combination of sensor technology. Both the wii-mote and the webcam are cheap but suffer some limitations in terms of tracking. By combining the two input methods, we may be able to overcome these.

Expected Outcomes:	A smart board interaction application which uses a combination of Wii and webcam to perform tracking.

Reference Material Links:	contact supervisor

Project Title:	Autonomous Robot for Search and Rescue

Name of Supervisor:	Claude Sammut

Email of Supervisor:	claude@cse.unsw.edu.au

Name of Joint/Co-Supervisor:	Bernhard Hengst

Email of Joint/Co-Supervisor:	bernhardh@cse.unsw.edu.au

School:	School of Computer Science and Engineering

For CSE and EET Projects:	School Project

Faculty Research Area (Theme):	Intelligent & Autonomous Systems

School Research Area:	Autonomous Systems & Sensing Technologies

Applicable to other Engineering schools/disciplines:

Abstract:	The aim of the project is to develop software for a robot designed for search and rescue missions. The software must enable the robot to explore its environment, creating a map and locating survivors. The robot, which is already operational, has a 3D camera, infrared and video cameras to enable it to navigate. There is a well-developed software infrastructure for basic control and tele-operation. The task for this project is to extend the software so that the robot can operate autonomously.

Research Environment:	The scholar will work with a team in the ARC Centre of Excellence for Autonomous Systems (CAS), including academic and technical staff and PhD students. CAS has a well-equipped rescue robot lab with several tracked vehicles and sophisticated sensor systems, as well as an aerial robot.

Novelty and Contribution:	The project will advance techniques in robot perception and navigation. It will contribute to the development of a practical robot for search and rescue. A successful outcome may be publishable in an international conference or journal.

Expected Outcomes:	The outcomes will include algorithms for perception and autonomous control of a rescue robot. The system will be evaluated in a RoboCup test arena in the CAS rescue robot laboratory.

Reference Material Links:	Sammut, C. Kadous, W., & Sheh, R. (2007). Learning to Drive Over Rough Terrain. In K. Furukawa (Ed.), In International Symposium on Skill Science, Tokyo. Sheh, R., Kadous, M.W., Sammut, C., & Hengst, B. (2007). Extracting Terrain Features from Range Images for Autonomous Random Stepfield Traversal. In D. Nardi (Ed.), IEEE International Conference on Safety, Security and Rescue Robotics, Rome. RoboCup Rescue we site: http://robotarenas.nist.gov/competitions.htm

Project Title:	Controlling virtual characters using artificial intelligence

Name of Supervisor:	Maurice Pagnucco

Email of Supervisor:	morri@cse.unsw.edu.au

Name of Joint/Co-Supervisor:

Email of Joint/Co-Supervisor:

School:	School of Computer Science and Engineering

For CSE and EET Projects:	School Project

Faculty Research Area (Theme):	Intelligent & Autonomous Systems

School Research Area:	Autonomous Systems & Sensing Technologies

Applicable to other Engineering schools/disciplines:	Mining Engineering

Abstract:	This project investigates the use of so-called cognitive robotics languages to control virtual characters in an interactive narrative. The project will be carried out in conjunction with the iCinema Centre for Interactive Cinema Research. In this project we will compare the use of cognitive robotics languages and agent programming languages for controlling animated virtual characters.

Research Environment:	This project will be carried out with the iCinema Centre for Interactive Cinema Research.

Novelty and Contribution:	The project is novel in a number of ways. Firstly it will compare two popular approaches to controlling intelligent agents. Secondly, it will look at the use of such approaches in specifying narrative goals that are to be achieved by the characters.

Expected Outcomes:	This project will involve developing software that will allow for the specification of narrative goals to be fulfilled by virtual characters in an interactive, immersive setting.

Reference Material Links:	John Funge. AI for Games and Animation: A Cognitive Modeling Approach. A K Peters, 1999

Project Title:	Localisation of Nao Robots on the Soccer Field

Name of Supervisor:	Bernhard Hengst

Email of Supervisor:	bernhardh@cse.unsw.du.au

Name of Joint/Co-Supervisor:	claude Sammut

Email of Joint/Co-Supervisor:	claude@cse.unsw.edu.au

School:	School of Computer Science and Engineering

For CSE and EET Projects:	School Project

Faculty Research Area (Theme):	Intelligent & Autonomous Systems

School Research Area:	Autonomous Systems & Sensing Technologies

Applicable to other Engineering schools/disciplines:

Abstract:	The success of the Robocup Standard Platform League soccer playing robots is critically dependent on their ability to determine their position, and the position of the ball on the soccer field. This localisation is performed by visual observation of the position and size of objects, such as goal-posts, field-lines, field carpet, other robots and the ball. The aim of the project is to track the position and orientation of a robot on the field from various geometric measurements associate with objects in the field-of-view and its own movement over time.

Research Environment:	The successful applicant will be working under the supervision of staff with many years of research and industrial experience. The research laboratory is located on the 3rd floor of the CSE building and houses students working on several related robotics projects. The successful candidate may be eligible to join the 2010 Robocup team.

Novelty and Contribution:	Cues for localisation in the league have become more challenging over the years, with the gradual removal of special localisation beacons. This research is intended to extract the maximum amount of information from the current soccer field. In particular, effective field-line localisation during the game using multiple hypothesis tracking is a new contribution. Good localisation is a prerequisite for successful play.

Expected Outcomes:	Real-time algorithms running on the Nao robot that can localise the robot accurately in terms of position on the soccer field and the direction it is facing using goal-post and field line information from vision.

Reference Material Links:	http://www.tzi.de/spl/bin/view/Website/WebHome http://cgi.cse.unsw.edu.au/~robocup/2008site/

Project Title:	Perceptual Anchoring in Mobile Robotics

Name of Supervisor:	Maurice Pagnucco

Email of Supervisor:	morri@cse.unsw.edu.au

Name of Joint/Co-Supervisor:	Claude Sammut

Email of Joint/Co-Supervisor:	claude@cse.unsw.edu.au

School:	School of Computer Science and Engineering

For CSE and EET Projects:	School Project

Faculty Research Area (Theme):	Intelligent & Autonomous Systems

School Research Area:	Autonomous Systems & Sensing Technologies

Applicable to other Engineering schools/disciplines:

Abstract:	Perceptual anchoring is the problem of relating (or anchoring) symbols to physical objects in the world via perception (e.g., vision) for the purpose of reasoning about a robot’s environment. In this project we will investigate one method proposed for solving the perceptual anchoring problem. We will use a Pioneer II robot to reason about objects in an office-like environment and use this knowledge to complete set tasks.

Research Environment:	This research project will be conducted in the ARC Centre of Excellence for Autonomous Systems lab at UNSW. The lab has several robotic platforms ranging from mobile Pioneer robots, through to robot arms, humanoid robots and rescue robots.

Novelty and Contribution:	Perceptual anchoring is a largely unsolved problem in Artificial Intelligence. This project has the potential to produce new and innovative research results.

Expected Outcomes:	Software running on Pioneer II under Player/Stage (playerstage.sf.net) and OpenCV capable of solving the perceptual anchoring problem for a limited class of pre-defined objects in an office-like environment.

Reference Material Links:	S. Coradeschi and A. Saffioti, “Perceptual Anchoring of Symbols for Actioin”, Proceeedings of the 17th IJCAI, pp. 407-412, 2001.

Project Title:	Programming a Humanoid Robot to Manipulate Complex Objects

Name of Supervisor:	Claude Sammut

Email of Supervisor:	claude@cse.unsw.edu.au

Name of Joint/Co-Supervisor:	Bernhard Hengst

Email of Joint/Co-Supervisor:	bernhardh@cse.unsw.edu.au

School:	School of Computer Science and Engineering

For CSE and EET Projects:	School Project

Faculty Research Area (Theme):	Intelligent & Autonomous Systems

School Research Area:	Autonomous Systems & Sensing Technologies

Applicable to other Engineering schools/disciplines:

Abstract:	One of the biggest challenges for robotics today is programming a robot to manipulate complex objects. One of the reasons for this is that it requires a combination of sensing with vision and touch in coordination with a motor system that can control a dextrous hand. The ARC Centre of Excellence for Autonomous Systems at UNSW has a lab with a humanoid robot with a pair of dextrous hands fitted with sophisticated tactile sensors. The robot also has a stereo camera for 3D vision. The aim of this project is to combine these sensors and program the robot to perform tasks that require both hands, e.g. assembling a child's toy.

Research Environment:	This project will be done with the assistance of the academic and technical staff of the ARC Centre of Excellence for Autonomous Systems (CAS). It will be conducted with with humanoid robot lab in the CSE building, which contains two robot arms and hands, assembled as a humanoid torso and a stereo camera system. All devices a programmed in C/C++.

Novelty and Contribution:	Despite many years of research in intelligent robotics, manipulation tasks, such as assembling an object from its components is still a difficult problem with many different approaches still to be explored.

Expected Outcomes:	It is expected that the scholar will write software for a demonstration of an assembly task such as described above.

Reference Material Links:	The main references are software and hardware manuals for the robot, which are available in the robotics lab.

Project Title:	RoboCup Humanoid Robot Soccer - Behaviours

Name of Supervisor:	Claude Sammut

Email of Supervisor:	claude@cse.unsw.edu.au

Name of Joint/Co-Supervisor:	Bernhard Hengst

Email of Joint/Co-Supervisor:	bernhardh@cse.unsw.edu.au

School:	School of Computer Science and Engineering

For CSE and EET Projects:	School Project

Faculty Research Area (Theme):	Intelligent & Autonomous Systems

School Research Area:	Autonomous Systems & Sensing Technologies

Applicable to other Engineering schools/disciplines:

Abstract:	UNSW has competed in the RoboCup robot soccer competition since 1999. We have been finalists in six of our nine years and world champions three times. In 2008, the competition introduced a new humanoid robot, called the Nao. This is a completely new platform and there is much to learn about how to program the robot to play soccer effectively. This project will look at behaviours for game play. These include locating and getting to the ball quickly; correct choice of behaviour in different parts of the field; attacker, supporter and goalie behavours; cooperation between team members; strategies to defeat opponents, etc.

Research Environment:	The scholar will work as part of a team developing the RoboCup code. We currently have nine Nao robots and a full scale field in the level 3 CSE robotics laboratory. The Nao's operating system is Linux, with additional control software provided by the manufacturer. There is also a freely available simulator for testing software. All code currently is written in C/C++.

Novelty and Contribution:	There are many different areas that the scholar may work on. These include using machine learning to improve the robot's locomotion; using novel high-level languages for planning and robot control; developing specialised behaviours for recovering from a fall, kicking goals, passing the ball, etc.

Expected Outcomes:	The scholar will develop software (and documentation) that extend the capabilities of the Nao robot and contribute to its game playing.

Reference Material Links:	See the reports of previous RoboCup teams at http://www.cse.unsw.edu.au/~robocup and also the main RoboCup competition site http://www.robocup.org

Project Title:	RoboCup Humanoid Robot Soccer - Locomotion

Name of Supervisor:	Claude Sammut

Email of Supervisor:	claude@cse.unsw.edu.au

Name of Joint/Co-Supervisor:	Bernhard Hengst

Email of Joint/Co-Supervisor:	bernhardh@cse.unsw.edu.au

School:	School of Computer Science and Engineering

For CSE and EET Projects:	School Project

Faculty Research Area (Theme):	Intelligent & Autonomous Systems

School Research Area:	Autonomous Systems & Sensing Technologies

Applicable to other Engineering schools/disciplines:

Abstract:	UNSW has competed in the RoboCup robot soccer competition since 1999. We have been finalists in six of our nine years and world champions three times. In 2008, the competition introduced a new humanoid robot, called the Nao. This is a completely new platform and there is much to learn about how to program the robot to play soccer effectively. This project will look at the bipedal locomotion that is fast, maneuverable and stable. Challenges including constructing gaits that can beat an opponent in a dash for the ball; agility in getting around the ball and dodging opponents; kicks for passing and scoring; fast methods for getting up from a fall; and knowing how to fall to avoid damage to the robot.

Research Environment:	The scholar will work as part of a team developing the RoboCup code. We currently have nine Nao robots and a full scale field in the level 3 CSE robotics laboratory. The Nao's operating system is Linux, with additional control software provided by the manufacturer. There is also a freely available simulator for testing software. All code currently is written in C/C++.

Novelty and Contribution:	There are many different areas that the scholar may work on. These include using machine learning to improve the robot's locomotion; using novel high-level languages for planning and robot control; developing specialised behaviours for recovering from a fall, kicking goals, passing the ball, etc.

Expected Outcomes:	The scholar will develop software (and documentation) that extend the capabilities of the Nao robot and contribute to its game playing.

Reference Material Links:	See the reports of previous RoboCup teams at http://www.cse.unsw.edu.au/~robocup and also the main RoboCup competition site http://www.robocup.org

Project Title:	Vision for RoboCup Nao Robots

Name of Supervisor:	Bernhard Hengst

Email of Supervisor:	bernhardh@cse.unsw.du.au

Name of Joint/Co-Supervisor:	Claude Sammut

Email of Joint/Co-Supervisor:	claude@cse.unsw.edu.au

School:	School of Computer Science and Engineering

For CSE and EET Projects:	School Project

Faculty Research Area (Theme):	Intelligent & Autonomous Systems

School Research Area:	Autonomous Systems & Sensing Technologies

Applicable to other Engineering schools/disciplines:

Abstract:	Vision is the primary sensor used by the soccer playing robots in the Robocup Standard Platform League. Vision is used to detect objects and other landmarks, such as goal-posts, field-lines, other robots and the ball. The aim of this project is to recognise and track visual objects reliably with the Nao robot as it moves about on the soccer field.

Research Environment:	The successful applicant will be working under the supervision of staff with many years of research and industrial experience. The research laboratory is located on the 3rd floor of the CSE building and houses students working on several related robotics projects. The successful candidate may be eligible to join the 2010 Robocup SPL team.

Novelty and Contribution:	With the gradual removal of special visual beacons, and the change in the characteristics of the robots, ie the move from four-legged to bipedal machines, new visual routines need to be developed. We wish to reliably recognize the new type of goals, field lines and other robots. There is significant opportunity to research and develop novel fast visual routines for this purpose. Visual skills are crucial for successful play and for autonomous robotics in general.

Expected Outcomes:	Real-time algorithms running on the Nao robot that can efficiently and reliably recognize visual objects and landmarks, to determine their various properties such as size and position relative to the robot.

Reference Material Links:	http://www.tzi.de/spl/bin/view/Website/WebHome http://cgi.cse.unsw.edu.au/~robocup/2008site/

Project Title:	Visual Odometry

Name of Supervisor:	Bernhard Hengst

Email of Supervisor:	bernhardh@cse.unsw.du.au

Name of Joint/Co-Supervisor:	Adam Milstein

Email of Joint/Co-Supervisor:	amilstein@cse.unsw.edu.au

School:	School of Computer Science and Engineering

For CSE and EET Projects:	School Project

Faculty Research Area (Theme):	Intelligent & Autonomous Systems

School Research Area:	Autonomous Systems & Sensing Technologies

Applicable to other Engineering schools/disciplines:	Sciences – Maths, Physics, Chemistry

Abstract:	Autonomous robots often slip when moving. The physical measurement of locomotion can be a poor indicator of their changing orientation. Visual Odometry is the process of estimating the movement of a (stereo) camera on a robot through its environment by matching point features between pairs of consecutive image frames. No prior knowledge of the scene is necessary. The aim of this project is to study the performance and reliability of visual odometry implemented on a real robot.

Research Environment:	The successful applicant will be working under the supervision of staff with many years of research and industrial experience. The research laboratory is located on the 3rd floor of the CSE building and houses students working on several related robotics projects.

Novelty and Contribution:	The research and implementation of reliable visual odometry from noisy sensor is challenging. A successful outcome would make a significant contribution in autonomous robotics for search & rescue, and soccer.

Expected Outcomes:	The development and implementation of real-time algorithms for estimating the movement of a camera through its environment by matching point features between pairs of consecutive image frames.

Reference Material Links:	Multiple-view Geometry, Richard Hartley and Andrew Zisserman. Visual Odometry Using Sparse Bundle Adjustment on an Autonomous Outdoor Vehicle, Niko Sueunderhauf, Kurt Konolige, Simon Lacroix, Peter Protzel.

Bioinformatics

Project Title:	iHMMune-align: machine learning models and antibody genes

Name of Supervisor:	Bruno Gaeta

Email of Supervisor:	bgaeta@unsw.edu.au

Name of Joint/Co-Supervisor:	Andrew Collins/Mike Bain

Email of Joint/Co-Supervisor:	a.collins@unsw.edu.au, mike@cse.unsw.edu.au

School:	School of Computer Science and Engineering

For CSE and EET Projects:	School Project

Faculty Research Area (Theme):	Health & Medical Technologies

School Research Area:	Bioinformatics

Applicable to other Engineering schools/disciplines:	Biomedical Engineering Sciences – Maths, Physics, Chemistry

Abstract:	This project is part of an effort to develop bioinformatics methods to understand the mechanisms through which the immune system is able to create antibodies against almost every possible bacteria/virus/parasite/toxin from a limited number of genes. We have previously developed iHMMune-align, a program that analyses DNA sequences encoding antibodies, which is now being adopted worldwide to study huge DNA sequence datasets in order to understand and improve treatment of conditions such as leukaemia and allergy. This project aims to improve the accuracy and robustness of iHMMune-align, as well as widening its applications. Good Java programming skills are required. Some understanding of the immune system and of bioinformatics would be useful but are not essential.

Research Environment:	The student will be working in the school of CSE and the school of Biotechnology and Biomolecular sciences, interacting with both computer scientists and immunologists.

Novelty and Contribution:

Expected Outcomes:	Improved antibody alignment software with broader applicability and robustness

Reference Material Links:	http://www.emi.unsw.edu.au/~ihmmune

Project Title:	Inferring regulatory networks from expression data

Name of Supervisor:	Bruno Gaeta

Email of Supervisor:	bgaeta@unsw.edu.au

Name of Joint/Co-Supervisor:	Mike Bain

Email of Joint/Co-Supervisor:	mike@cse.unsw.edu.au

School:	School of Computer Science and Engineering

For CSE and EET Projects:	School Project

Faculty Research Area (Theme):	Health & Medical Technologies

School Research Area:	Bioinformatics

Applicable to other Engineering schools/disciplines:

Abstract:	We have previously developed MINER (Microarray Interactive Network Exploration and Representation), a bioinformatics application that allows predicting potential gene regulatory relationships from gene expression data, exploring these predicted relationships in an interactive manner and combining these into potential gene regulatory networks. MINER has been used successfully to identify new genes of interest in the development of some cancers. The goal of this project is to improve MINER by incorporating new machine learning algorithms and streamlining the interface to make the program more widely useable and robust. This project is best suited to students with some experience with machine learning and PHP programming.

Research Environment:	The student will be working in the school of CSE, interacting with both supervisors and with other students working on the project. If the project goes well the student will also be working with end users of the software (biologists)

Novelty and Contribution:

Expected Outcomes:	New algorithms for analysing gene expression data and web-based software using these algorithms

Reference Material Links:	Contact supervisor

Databases

Project Title:	An empirical evaluation of similarity search techniques in time series data

Name of Supervisor:	Ghazi Al-Naymat

Email of Supervisor:	ghazi@cse.unsw.edu.au

Name of Joint/Co-Supervisor:	Sherif Sakr

Email of Joint/Co-Supervisor:	ssakr@cse.unsw.edu.au

School:	School of Computer Science and Engineering

For CSE and EET Projects:	School Project

Faculty Research Area (Theme):	eResearch (Knowledge and Services Engineering)

School Research Area:	Databases

Applicable to other Engineering schools/disciplines:

Abstract:	Time Series are ubiquitous; hence, similarity search is one of the biggest challenges in the area of mining time series data. This is due to the vast data size -- number of sequences and number of dimensions that lead to a very costly querying process. In this empirical study, we aim to evaluate the state-of-the-art techniques that have been used in time series similarity searches. Two different similarity measures will be used in this study, namely Dynamic Time Warping (DTW) and Euclidean distance. The evaluation focuses in showing the techniques accuracy as well as their performance by conducting thorough and exhaustive experiments on synthetic and real datasets.

Research Environment:	The student will work in a team of researchers. A literature review is required to understand survey the previous work done in the area of mining time series data. Implementing some of the similarity techniques and conducting comprehensive experiments are required.

Novelty and Contribution:	An empirical evaluation to the state-of-the-art similarity search techniques in time series data

Expected Outcomes:	A thorough study that shows a comprehensive evaluation for the latest similarity search in time series data using two measures – DTW and Euclidean distance. A technical report that will be publishable as a conference or journal paper.

Reference Material Links:	Keogh, E. J. & Kasetty, S. On the need for time series data mining benchmarks: a survey and empirical demonstration. KDD, 2002, 102-111 Ghazi Al-Naymat and Javid Taheri. Effects of Dimensionality Reduction Techniques on Time Series Similarity Measurements. Proceedings of the 6th ACS/IEEE International Conference on Computer Systems and Applications (AICCSA-08), Doha, Qatar. Mar 31st- Apr 4th, 2008. Pages (188-196). Sakoe, H. and Chiba, S., Dynamic programming algorithm optimization for spoken word recognition, IEEE Transactions on Acoustics, Speech and Signal Processing, 26(1) pp. 43- 49, 1978, ISSN: 0096-3518 Han, J. & Kamber, M. Data Mining: Concepts and Techniques. The Morgan Kaufmann, 2001

Project Title:	An Implementation of SPARQL Query Processor

Name of Supervisor:	Sherif Sakr

Email of Supervisor:	ssakr@cse.unsw.edu.au

Name of Joint/Co-Supervisor:

Email of Joint/Co-Supervisor:

School:	School of Computer Science and Engineering

For CSE and EET Projects:	School Project

Faculty Research Area (Theme):	Programming Languages and Software Engineering

School Research Area:	Databases

Applicable to other Engineering schools/disciplines:

Abstract:	The vision of the Semantic Web has brought about new challenges at the intersection of web research and data management. One fundamental research issue at this intersection is the storage of the Resource Description Framework (RDF) data. The RDF data model has been designed as a exible representation of schema-relaxable or even schema-free information. In RDF, all data items are represented in the form of (subject, predicate, object) triples, also known as (subject, property, value) triples. The target of the project is to implement an approach for storing and querying RDF data using the robust infrastructure of relational database management systems. The idea of the project is to avoid using an extreme vertical fragmentation of the RDF data or an extreme horizontal triple storage of the RDF data. However, the focus is to utilize the fact that each RDF dataset requires a tailored table schema that achieves ecient query processing based on the query workload. Relational optimization techniques will be used to reduce the number of joins operations in the query plan. Moreover, the awareness of the query workload information will be used to enhance the vertical fragmentation of RDF data by keeping null storage below a given threshold.

Research Environment:	The student will work in an international group of PhD students, researchers and senior researchers in the Service Oriented Computing Research Group. Some literature review will be required to learn the basics of RDF and SPARQL query processing. Experimental analysis skills will also be acquired during the project activities. Project results will have a very good chance to be published in a good venue.

Novelty and Contribution:	- Providing a flexible storage schema of RDF data based on query workloads. - Ecient query processing of RDF data using the robust infrastructure of relational database management systems.

Expected Outcomes:	- This project will involve developing of an ecient query processor for RDF and SPARQL queries. - Literature scan of RDF storing and querying bibliography. - Comparison between the proposed RDF query processor and the other proposed techniques.

Reference Material Links:	- RDF Processing Bibliography: http://www.cse.unsw.edu.au/ssakr/RDFBiblio.htm - An introduction to RDF and SPARQL http://www.dajobe.org/talks/200603-sparql-stanford/ http://research.talis.com/2005/rdf-intro/ http://www.rdfabout.com/quickintro.xpd

Project Title:	Mining pairs trading patterns from a large stock market data

Name of Supervisor:	Ghazi Al-Naymat

Email of Supervisor:	ghazi@cse.unsw.edu.au

Name of Joint/Co-Supervisor:	Sherif Sakr

Email of Joint/Co-Supervisor:	ssakr@cse.unsw.edu.au

School:	School of Computer Science and Engineering

For CSE and EET Projects:	School Project

Faculty Research Area (Theme):	eResearch (Knowledge and Services Engineering)

School Research Area:	Databases

Applicable to other Engineering schools/disciplines:

Abstract:	The volume of financial data has rapidly increased due to advances in software and hardware technologies. Stock markets provide examples of financial data that contains many attributes – far more than traders can readily understand. Traders nonetheless attempt to determine relationships between data attributes that can yield profitable trading of financial instruments. As traders’ needs have become more complex, the demand for more efficient techniques has grown. Many researchers have developed algorithms and frameworks that concentrate on mining useful patterns in stock market datasets such pairs trading. The literature has shown that pairs trading pattern is one of the most sought-after patterns because of its market-neutral strategy -- the return is uncorrelated to the market. This project aims to design a model, which can capture pairs trading patterns from large stock market data. The devised model should incorporate well-known similarity measure; that is Dynamic Time Warping (DTW).

Research Environment:	The student will work in a team of researchers. A literature review is required to understand survey the previous work done in the area of mining stock market data. Implementing the proposed model and conducting comprehensive experiments are required.

Novelty and Contribution:	Devising an efficient algorithm to find pairs trading patterns from large stock market data

Expected Outcomes:	Novel algorithm to discover pairs trading patterns from large stock market data. A technical report that will be publishable as a conference or journal paper.

Reference Material Links:	Granapathy Vidyamurthy. Pairs Trading Quantitative Methods and Analysis. Wiley, 2004. Longbing Cao, Dan Luo, and Chengqi Zhang. Fuzzy genetic algorithms for pairs mining. In Proceedings of the 9th Paci?c Rim International Conference on Arti?cial Intelli- gence (PRICAI), volume 4099 of Lecture Notes in Computer Science, pages 711–720. Springer Berlin / Heidelberg, 2006b. Sakoe, H. and Chiba, S., Dynamic programming algorithm optimization for spoken word recognition, IEEE Transactions on Acoustics, Speech and Signal Processing, 26(1) pp. 43- 49, 1978, ISSN: 0096-3518 Han, J. & Kamber, M. Data Mining: Concepts and Techniques. The Morgan Kaufmann, 2001

Project Title:	ORM vs JDBC: Do we still need SQL?

Name of Supervisor:	John Shepherd

Email of Supervisor:	jas@cse.unsw.edu.au

Name of Joint/Co-Supervisor:	Helen Paik

Email of Joint/Co-Supervisor:	hpaik@cse.unsw.edu.au

School:	School of Computer Science and Engineering

For CSE and EET Projects:	School Project

Faculty Research Area (Theme):	eResearch (Knowledge and Services Engineering)

School Research Area:	Databases

Applicable to other Engineering schools/disciplines:

Abstract:	In the early 2000's people built Web/database interfaces by specifying database access in SQL and application logic in a programming language such as PHP or Java. More recently, a range of frameworks (mainly Java) have been proposed which allow programmers to model data at the program/object level and have SQL to access the database (semi)automatically generated via a "persistence mapping framework". Given the "semantic gap" between tuples and objects and the different ways that tuples and objects are accessed, it seems that there might be a performance penalty in using such an approach to build Web/database applications. This project aims to investigate the relative efficiency of database access generated via various persistence mappings as opposed to "hand-crafted" SQL interfaces between databases and programming languages, and contrast this with the productivity gains claimed for such frameworks. This project will build on work from last summer which suggested that ORMs were significantly less efficient in some situations; these situations need to be characterised and precise performance benefits/losses determined.

Research Environment:	Working with John Shepherd and Helen Paik using a range of tools on the CSE servers (including PostgreSQL, SQLite, Java, JDBC, Tomcat, Hibernate, Ruby on Rails). The plan is to take some existing applications and re-cast them in an alternative framework and then do some experimental performance analysis to determine the relative efficiencies.

Novelty and Contribution:	This project aims to answer an as-yet unexplored aspect of the tools used to develop Web/database applications.

Expected Outcomes:	Empirical evidence and analysis that will help to provide some insight into the question: "Do persistence frameworks render hand-crafted database access solutions obsolete?". If the answer is not clear-cut, then we also hope to characterise what kinds of applications are problematic for each approach.

Reference Material Links:	"Integrating Programming Languages and Database: What's the Problem?", William R. Cook, Ali H. Ibrahim, available via: http://www.odbms.org/experts.html#article10 Also, typing key phrases like "object-relational mapping" and "active records" and "persistence framework" into Google will lead to a plethora of resources.

Embedded, Real Time & Operating Systems

Project Title:	Audio Framework for Embedded OS

Name of Supervisor:	Ihor Kuz

Email of Supervisor:	Ihor.kuz@nicta.com.au

Name of Joint/Co-Supervisor:

Email of Joint/Co-Supervisor:

School:	School of Computer Science and Engineering

For CSE and EET Projects:	NICTA Project

Faculty Research Area (Theme):	Embedded Systems and Communications

School Research Area:	Embedded, Real Time & Operating Systems

Applicable to other Engineering schools/disciplines:

Abstract:	Multimedia and audio devices are a popular class of embedded systems. As part of our research into modularised, (seL4 or OKL4) microkernel-based operating systems for embedded devices, we wish to look at the design and implementation of an audio framework. This involves designing and developing a reusable software framework for audio applications and devices. The framework must be built using the component architecture we have developed for microkernel-based operating systems. Building a demonstrator showing the framework in use will also be part of the work.

Research Environment:	The project will be conducted with the ERTOS group, one of the world's leading operating-systems research groups, in a stimulating environment that combines cutting-edge research with building systems that are deployed in real products.

Novelty and Contribution:	This project will provide the foundation for building future audio-enabled embedded systems. Another important result will be the further evaluation and improvement of our component architecture, and input into research regarding structure and performance of componentised operating systems.

Expected Outcomes:	A well designed, componentised, audio framework, and a demonstrator system that implements an audio-enabled system based on this framework.

Reference Material Links:	ERTOS: http://www.ertos.nicta.com.au/ seL4: http://www.ertos.nicta.com.au/research/sel4/ L4: http://portal.ok-labs.com/ http://www.l4hq.org/ http://www.l4ka.org/ Course Prerequisites: Operating Systems. Advanced OS is highly recommended Skill Prerequisites: A good knowledge of C and system-level programming are necessary.

Project Title:	Covert Channels - building infrastructure to steal secrets.

Name of Supervisor:	Kevin Elphinstone

Email of Supervisor:	kevine@cse.unsw.edu.au

Name of Joint/Co-Supervisor:

Email of Joint/Co-Supervisor:

School:	School of Computer Science and Engineering

For CSE and EET Projects:	NICTA Project

Faculty Research Area (Theme):	Embedded Systems and Communications

School Research Area:	Embedded, Real Time & Operating Systems

Applicable to other Engineering schools/disciplines:	Electrical Engineering & Telecommunications

Abstract:	Many operating systems are explicitly designed to keep secrets exactly that - secret. They actively seek to prevent data from leaking from one data clasification (e.g. top secret) to another lower clasification - even if the user actively attempts to violate system security. Your mission, should you choose to accept it, is to develop software to bypass system security to leak secrets using covert channels. The project would be fairly open for the student to survey communications techniques includling signal processing and error correction to develop tools capable of sending secret messages to a collaborator.

Research Environment:	The project will be conducted within the ERTOS group at NICTA at Kensington (L5 building). ERTOS is one of the world's leading operating-systems research groups. It is a stimulating environment that combines cutting-edge research with building systems that are deployed in real products. ERTOS is a team environment that enables and values contributions from all its members.

Novelty and Contribution:

Expected Outcomes:

Reference Material Links:

Project Title:	Embedded File System

Name of Supervisor:	Ihor Kuz

Email of Supervisor:	Ihor.kuz@nicta.com.au

Name of Joint/Co-Supervisor:

Email of Joint/Co-Supervisor:

School:	School of Computer Science and Engineering

For CSE and EET Projects:	NICTA Project

Faculty Research Area (Theme):	Embedded Systems and Communications

School Research Area:	Embedded, Real Time & Operating Systems

Applicable to other Engineering schools/disciplines:

Abstract:	One of the key services that an OS provides is a file system. We are in the midsts of designing and building a modular (OKL4 or seL4) microkernel-based operating system for embedded devices. While there are many file-systems available, we do not yet have a suitable file system service for our OS. Furthermore, not much work has been done on the design of a file system service in a componentised environment. This project involves designing and implementing an existing filesystem to work in a componentised operating system. Besides providing a functioning and reusable system component, it is necessary that the resulting file system also exhibits good performance.

Research Environment:	The project will be conducted with the ERTOS group, one of the world's leading operating-systems research groups, in a stimulating environment that combines cutting-edge research with building systems that are deployed in real products.

Novelty and Contribution:	Besides contributing to the development of an embedded OS, this project will provide insights into the design issues and performance tradeoffs involved with componentising system services.

Expected Outcomes:	A working file system integrated into our componentised OS. The insight into the design decisions and tradeoffs involved in designing such a subsystem will influence and further drive research into structuring modular, microkernel- based operating systems.

Reference Material Links:	ERTOS: http://www.ertos.nicta.com.au seL4: http://www.ertos.nicta.com.au/research/sel4/ L4: http://portal.ok-labs.com http://www.l4hq.org/ http://www.l4ka.org/ Course Prerequisites: Operating Systems. Advanced OS is highly recommended Skill Prerequisites: A good knowledge of C and system-level programming are necessary.

Project Title:	Event-Based Device Drivers

Name of Supervisor:	Peter Chubb

Email of Supervisor:	peter.chubb@nicta.com.au

Name of Joint/Co-Supervisor:	Leonid Ryzhyk

Email of Joint/Co-Supervisor:	leonid.ryzhyk@nicta.com.au

School:	School of Computer Science and Engineering

For CSE and EET Projects:	NICTA Project

Faculty Research Area (Theme):	Embedded Systems and Communications

School Research Area:	Embedded, Real Time & Operating Systems

Applicable to other Engineering schools/disciplines:

Abstract:	The Dingo device driver architecture, developed at ERTOS, helps driver writers produce better code, free of many types of errors found in conventional drivers, including concurrency and protocol errors. The initial version of Dingo was successful in demonstrating the effectiveness of the approach. In order to facilitate wider adoption of the Dingo architecture, we are currently building an open source version of the Dingo framework for Linux. This project's aim is to develop several sample drivers for Dingo. These drivers will be used to test and benchmark the framework and will also serve as examples demonstrating advantages of the Dingo architectures to the Linux kernel community.

Research Environment:	The project will be conducted with the ERTOS group, one of the world's leading operating-systems research groups, in a stimulating environment that combines cutting-edge research with building systems that are deployed in real products.

Novelty and Contribution:

Expected Outcomes:	One or more working drivers for common devices such as Ethernet and audio controllers.

Reference Material Links:	A description of the Dingo device driver architecture http://ertos.nicta.com.au/publications/papers/Ryzhyk_CKH_09.pdf Course Prerequisites: Operating Systems. Advanced OS is highly recommended Skill Prerequisites: A good knowledge of C and system-level programming are necessary.

Project Title:	Interrupt handling for realtime systems.

Name of Supervisor:	Peter Chubb

Email of Supervisor:	peter.chubb@nicta.com.au

Name of Joint/Co-Supervisor:

Email of Joint/Co-Supervisor:

School:	School of Computer Science and Engineering

For CSE and EET Projects:	NICTA Project

Faculty Research Area (Theme):	Embedded Systems and Communications

School Research Area:	Embedded, Real Time & Operating Systems

Applicable to other Engineering schools/disciplines:

Abstract:	The standard Unix/Linux interrupt handler runs at a very high priority and steals time from the currently running process. As such, under heavy interrupt loads, real time processes can miss their deadlines. There is currently a patch available for Linux, called PREEMPT:RT, that moves interrupt handling to threads that can be given a lower priority than real time processes. However, it seems from the preliminary results that we have as if the advantages possible from using a thread per interrupt are not being fully utilised, and there is a significant performance hit. This project is to evaluate the current, and PREEMPT:RT interrupt handling styles, and to devise more efficient ways to handle interrupts.

Research Environment:	The student will be working in a team with PhD and Masters students. He or she will also be expected to contribute to and be part of the Open Source Linux Kernel community.

Novelty and Contribution:	Threaded interrupt handlers are not new: they are used in, e.g, FreeBSD. However, the user level driver work we have done suggests that a threaded interrupt handler can be made to have better system throughput (at the expense of some amount of latency) than a traditional model. This work is to confirm or deny the performance possibilities, by careful measurement after some implementation.

Expected Outcomes:	We expect that at least one device driver can be modified to give better performance while not interfering with real time tasks. The work should be reportable in Linux conferences.

Reference Material Links:	Contact supervisor

Project Title:	Linux as a Boot Loader

Name of Supervisor:	Peter Chubb

Email of Supervisor:	peter.chubb@nicta.com.au

Name of Joint/Co-Supervisor:

Email of Joint/Co-Supervisor:

School:	School of Computer Science and Engineering

For CSE and EET Projects:	NICTA Project

Faculty Research Area (Theme):	Embedded Systems and Communications

School Research Area:	Embedded, Real Time & Operating Systems

Applicable to other Engineering schools/disciplines:

Abstract:	Linux has a feature called `KEXEC' that allows a linux kernel to replace itself with another kernel. Therefore, it should be possible to use a small Linux instance as a boot loader, to boot over NFS or TFTP, without some of the problems that other boot loaders (e.g., U-boot) have. In fact this is done in a small way for the Sharp Zaurus Angstrom distribution. However, it would be really nice to be able to boot systems other than Linux. This project is to develop a root filesystem that in the first case can be used to boot another Linux instance on the BeagleBoard, and when successful, to be able to boot SeL4 plus other components as required.

Research Environment:	The project will be conducted with the ERTOS group, one of the world's leading operating-systems research groups, in a stimulating environment that combines cutting-edge research with building systems that are deployed in real products. Students will work with PhD and Masters students, as well as with researchers and research engineers.

Novelty and Contribution:	Using Linux to boot Linux is not new, however, using it to boot a different operating system has not previously been done.

Expected Outcomes:	We envisage that if successful this project will allow the use of Linux userspace as a pre-boot environment for booting the SeL4 microkernel on an embedded system.

Reference Material Links:	OpenEmbedded: http://www.openembedded.org SeL4: http://ertos.nicta.com.au/research/sel4/ BeagleBoard: http://beagleboard.org

Project Title:	Linux as a Component

Name of Supervisor:	Ihor Kuz

Email of Supervisor:	Ihor.kuz@nicta.com.au

Name of Joint/Co-Supervisor:

Email of Joint/Co-Supervisor:

School:	School of Computer Science and Engineering

For CSE and EET Projects:	NICTA Project

Faculty Research Area (Theme):	Embedded Systems and Communications

School Research Area:	Embedded, Real Time & Operating Systems

Applicable to other Engineering schools/disciplines:

Abstract:	The ERTOS group has done (and commercialised) much work in virtualising Linux to run on the OKL4 and seL4 microkernels. We have also done work developing a componentised microkernel-based OS. However the two essentially live in separate worlds. The goal of this project is to integrate virtualised Linux (and its applications) into the componentised OS. One way to do this is to treat Linux as a large component and develop appropriate interfaces and an appropriate framework for this. The project will investigate the best way to do this and implement a prototype system.

Research Environment:	The project will be conducted with the ERTOS group, one of the world's leading operating-systems research groups, in a stimulating environment that combines cutting-edge research with building systems that are deployed in real products.

Novelty and Contribution:	This project will combine two aspects of the novel research that we do at ERTOS, making both more useful and avaliable to a broader audience of developers. It will lead to a system that can be used in a variety of existing and future embedded devices, including mobile phones and media players.

Expected Outcomes:	A demonstration system consisting of a componentised OS, a virtualised Linux instance, and applications all coexisting and communicating in harmony.

Reference Material Links:	ERTOS: http://www.ertos.nicta.com.au/ seL4: http://www.ertos.nicta.com.au/research/sel4/ L4 and virtualised Linux: http://portal.ok-labs.com/ Componentised OS: http://www.ertos.nicta.com.au/research/camkes/ Course Prerequisites: Operating Systems. Advanced OS is highly recommended Skill Prerequisites: A good knowledge of C and system-level programming are necessary.

Project Title:	Locking tradeoffs for multiprocessor kernels

Name of Supervisor:	Gernot Heiser

Email of Supervisor:	gernot@unsw.edu.au

Name of Joint/Co-Supervisor:	Kevin Elphinstone

Email of Joint/Co-Supervisor:	kevine@cse.unsw.edu.au

School:	School of Computer Science and Engineering

For CSE and EET Projects:	NICTA Project

Faculty Research Area (Theme):	Embedded Systems and Communications

School Research Area:	Embedded, Real Time & Operating Systems

Applicable to other Engineering schools/disciplines:	Electrical Engineering & Telecommunications

Abstract:	Multi-threaded (SMT) processors are characterised by shared caches and very low (single-cycle) communication latencies between execution contexts. Consequently, treads on such a system are scheduled from a single scheduling queue, and other kernel data structures are also shared. This makes fine-grained locking hard and potentially expensive. For a kernel where average latencies of kernel operations are very short, a global kernel lock (i.e., single-threaded kernel) could be an appropriate approach. This project is to investigate global vs fine-grained kernel locking on the OKL4 microkernel running on a highly (>4) multi-threaded processor core. This is likely to lead to publishable results.

Research Environment:	Operating systems (ERTOS) group at NICTA NRL

Novelty and Contribution:	Re-assessment of an old issue with new constraints: microkernel with very short in-kernel execution paths and modern multi-threaded hardware

Expected Outcomes:	Research report analysing tradeoffs and quantifying their performance implications

Reference Material Links:	http:ertos/nicta.com/au

Project Title:	Lottery Scheduling for Embedded Systems

Name of Supervisor:	Kevin Elphinstone

Email of Supervisor:	kevine@cse.unsw.edu.au

Name of Joint/Co-Supervisor:

Email of Joint/Co-Supervisor:

School:	School of Computer Science and Engineering

For CSE and EET Projects:	NICTA Project

Faculty Research Area (Theme):	Embedded Systems and Communications

School Research Area:	Embedded, Real Time & Operating Systems

Applicable to other Engineering schools/disciplines:	Electrical Engineering & Telecommunications

Abstract:	A flexible CPU scheduling mechanism should be able to accommodate many application domains with only configuration of the scheduler being required. One such scheduler that might fulfill the role is a lottery scheduler. A lottery scheduler randomly draws a winner to determine which task runs next. The distribution of tickets to tasks is used to affect the scheduling behaviour. This project aims to develop an efficient lottery scheduler and an API for managing the ticket distribution.

Research Environment:	The project will be conducted within the ERTOS group at NICTA at Kensington (L5 building). ERTOS is one of the world's leading operating-systems research groups. It is a stimulating environment that combines cutting-edge research with building systems that are deployed in real products. ERTOS is a team environment that enables and values contributions from all its members.

Novelty and Contribution:

Expected Outcomes:

Reference Material Links:

Project Title:	Multi-VDD Power Analysis

Name of Supervisor:	Sri Parmeswaran

Email of Supervisor:	sridevan@cse.unw.edu.au

Name of Joint/Co-Supervisor:	Jorgen Peddersen

Email of Joint/Co-Supervisor:	jorgenp@cse.unsw.edu.au

School:	School of Computer Science and Engineering

For CSE and EET Projects:	School Project

Faculty Research Area (Theme):	Embedded Systems and Communications

School Research Area:	Embedded, Real Time & Operating Systems

Applicable to other Engineering schools/disciplines:

Abstract:	This aim of this project is to analyse the affect of run-time dynamic alteration of the supply voltage in an embedded system. The student will work with Synopsys tools to calculate the affect of and power savings achieved by adjusting the voltage of an existing processor. Power consumption of an embedded system can be drastically reduced by changing the input voltage. However, the maximum clocking frequency also lowers, reducing performance of the device. Working out the times and costs involved to optimise power consumption is a major research field, although the operating environment typically needs to be known beforehand. The proposed project will allow dynamic monitoring of the system to be performed to allow a system to optimise power consumption at run-time without prior knowledge of the environment.

Research Environment:	The student will be working with the aid of a post-doc researcher who designed the processor to be analysed.

Novelty and Contribution:	Run-time optimisation of power is a relatively new field and the techniques created in this project will be applicable to other ongoing research in the field.

Expected Outcomes:	It is expected to achieve an accurate model of the affects of dynamic voltage scaling on the provided processor and to give that processor the ability to make scaling decisions in parallel with application execution.

Reference Material Links:	http://ieeexplore.ieee.org/xpls/abs_all.jsp?arnumber=4196148 http://ieeexplore.ieee.org/xpls/abs_all.jsp?arnumber=4092075

Project Title:	Networks for Reconfigurable Systems on a Chip

Name of Supervisor:	Oliver Diessel

Email of Supervisor:	odiessel@cse.unsw.edu.au

Name of Joint/Co-Supervisor:

Email of Joint/Co-Supervisor:

School:	School of Computer Science and Engineering

For CSE and EET Projects:	School Project

Faculty Research Area (Theme):	Embedded Systems and Communications

School Research Area:	Embedded, Real Time & Operating Systems

Applicable to other Engineering schools/disciplines:	Electrical Engineering & Telecommunications

Abstract:	Networks on chip for reconfigurable systems on chip need to be tuned according to prevailing application requirements. This need creates a demand for new approaches and methods for dynamically optimizing on-chip communication structures. In this ToR project you will build on your knowledge of digital systems, FPGA and network technology to assist a small, diverse team of researchers develop a simple demonstration of on-the-fly network-on-chip adaptation.

Research Environment:	You will work with a team comprising two masters students and the supervisor using a variety of FPGA boards and design workstations available in the School. The International Conference on Field-Programmable Technology will be hosted by the University during the project period. You will be encouraged to assist and participate in the proceedings.

Novelty and Contribution:	The project provides the opportunity to quickly become exposed to and participate in leading edge reconfigurable systems research. The project serves the valuable purposes of tying together several threads of research currently under exploration, and providing a proof of concept necessary to obtain wider support for the work.

Expected Outcomes:	A demonstration of the efficient reconfiguration of an on-chip communication network necessary to maintain quality of service when application requirements change.

Reference Material Links:	NoCs for reconfigurable SoCs: * Marescaux et al., Run-time support for heterogeneous multitasking on reconfigurable SoCs * Bjerregaard et al., A survey of research and practices of network-on-chip * Warkentin, NoC support for dynamic FPGA pages Dynamic reconfiguration of FPGAs: * Ullmann et al., An FPGA run-time system for dynamical on-demand reconfiguration * Lai, A new API for internal reconfiguration of Xilinx FPGAs

Project Title:	NoTA prototype on OKL4

Name of Supervisor:	Gernot Heiser

Email of Supervisor:	gernot@cse.unsw.edu.au

Name of Joint/Co-Supervisor:	Ihor Kuz

Email of Joint/Co-Supervisor:	ihor.kuz@nicta.com.au

School:	School of Computer Science and Engineering

For CSE and EET Projects:	NICTA Project

Faculty Research Area (Theme):	Embedded Systems and Communications

School Research Area:	Embedded, Real Time & Operating Systems

Applicable to other Engineering schools/disciplines:

Abstract:	The Network on Terminal Architecture (NoTA) is an emerging approach for structuring software on a mobile phone handset, turning it logically into a distributed system. OKL4 should be the ideal platform for supporting NoTA with low overhead. This project is to demonstrate this by building a prototype NoTA-based system, and evaluate its performance.

Research Environment:	The project will be conducted with the NICTA ERTOS group, one of the world's leading operating-systems research groups, in a stimulating environment that combines cutting-edge research with building systems that are deployed in real products.

Novelty and Contribution:

Expected Outcomes:

Reference Material Links:	NoTa: http://en.wikipedia.org/wiki/Network_on_Terminal_Architecture OKL4: http://okl4.org/ Course Prerequisites: Operating Systems. Advanced OS is highly recommended Skill Prerequisites: A good knowledge of C and system-level programming are necessary.

Project Title:	Optimising memory management in a high-performance secure microkernel

Name of Supervisor:	Kevin Elphinstone

Email of Supervisor:	kevine@cse.unsw.edu.au

Name of Joint/Co-Supervisor:

Email of Joint/Co-Supervisor:

School:	School of Computer Science and Engineering

For CSE and EET Projects:	NICTA Project

Faculty Research Area (Theme):	Embedded Systems and Communications

School Research Area:	Embedded, Real Time & Operating Systems

Applicable to other Engineering schools/disciplines:	Electrical Engineering & Telecommunications

Abstract:	The seL4 microkernel is designed to give unprecedented reliability and security via a formal proof of correctness. However, up until now, seL4's design has focused on correctness, amenability to proof, and security. Limited attention has been given to optimising memory management. This project involves identifying, proposing, and exploring solutions to memory management issues in a modern microkernel. It will give the opportunity for a clever student to have real impact on the evolution of leading-edge operating system technology.

Research Environment:	The project will be conducted within the ERTOS group at NICTA at Kensington (L5 building). ERTOS is one of the world's leading operating-systems research groups. It is a stimulating environment that combines cutting-edge research with building systems that are deployed in real products. ERTOS is a team environment that enables and values contributions from all its members.

Novelty and Contribution:

Expected Outcomes:

Reference Material Links:

Project Title:	Performance limits of IPC fastpath implemented in C

Name of Supervisor:	Gernot Heiser

Email of Supervisor:	gernot@unsw.edu.au

Name of Joint/Co-Supervisor:	Kevin Elphinstone

Email of Joint/Co-Supervisor:	kevine@cse.unsw.edu.au

School:	School of Computer Science and Engineering

For CSE and EET Projects:	NICTA Project

Faculty Research Area (Theme):	Embedded Systems and Communications

School Research Area:	Embedded, Real Time & Operating Systems

Applicable to other Engineering schools/disciplines:	Electrical Engineering & Telecommunications

Abstract:	High-performance L4 microkernels use traditionally an assembler “fastpath” to overcome the performance problems resulting from C code in the critical IPC system call path. Relying on assembler code has a high engineering and maintenance cost, and makes formal verification harder. Eliminating the need for assembler code, or, at least, reducing the amount needed, would have significant practical benefits. This thesis is to analyse the IPC path in order to understand why C compilers do not do a better job on it. It will analyse possible (semantically-invariant) modifications of the C code, in order to investigate how far C performance can be pushed, and whether it can be made competitive with the assembler implementation. This study is to be done on at least the ARM architecture (x86 optional) and using several compilers (gcc, RVCT, maybe Green Hills).

Research Environment:	Operating systems (ERTOS) group at NICTA NRL

Novelty and Contribution:	Thorough investigation of performance limits of C code for low-level high-erformance OS kernel code

Expected Outcomes:	A thorough analysis of the code and its optimisability. Potentially publishable.

Reference Material Links:	http:ertos/nicta.com/au

Project Title:	Power Analysis of Leon3 Processor

Name of Supervisor:	Sri Parameswaran

Email of Supervisor:	sridevan@cse.unsw.edu.au

Name of Joint/Co-Supervisor:	Jorgen Peddersen

Email of Joint/Co-Supervisor:	jorgenp@cse.unsw.edu.au

School:	School of Computer Science and Engineering

For CSE and EET Projects:	School Project

Faculty Research Area (Theme):	Embedded Systems and Communications

School Research Area:	Embedded, Real Time & Operating Systems

Applicable to other Engineering schools/disciplines:

Abstract:	The aim of the project is to aid the production of a power model of the LEON 3 processor designed by Gaisler Research. The student will use Synopsys tools to simulate and analyse the model to determine events that consume power and attempt to produce a simple model of that power consumption using macro modelling techniques. The project also involves use of mathematical tools to create the model based on the power consuming events.

Research Environment:	The student will work with the aid of a post-doc researcher who created a similar model for a different processor.

Novelty and Contribution:	The power model created will be useful to the research community and has several uses for integration into power optimisation systems.

Expected Outcomes:	The creation of this power model will allow the design of a hybrid processor that can estimate its own power consumption in parallel to runtime.

Reference Material Links:	Leon 3 reference: http://www.gaisler.com/cms/index.php?option=com_content&task=view&id=13&Itemid=53

Project Title:	seL4 vs Singularity

Name of Supervisor:	Gernot Heiser

Email of Supervisor:	gernot@unsw.edu.au

Name of Joint/Co-Supervisor:	Kevin Elphinstone

Email of Joint/Co-Supervisor:	kevine@cse.unsw.edu.au

School:	School of Computer Science and Engineering

For CSE and EET Projects:	NICTA Project

Faculty Research Area (Theme):	Embedded Systems and Communications

School Research Area:	Embedded, Real Time & Operating Systems

Applicable to other Engineering schools/disciplines:

Abstract:	Singularity (using language-based protection) and OKL4/seL4 (using hardware-based protection) are the leading examples of two alternative approaches to OS kernels for high security. This project is to do a quantitative comparison of the two approaches, focusing on the size of the trusted computing base and the performance of systems buillt on top (starting with assessing the basic communication performance).

Research Environment:	Operating systems (ERTOS) group at NICTA NRL

Novelty and Contribution:	Quantitiive comparison of two basic approaches to secure operating systems

Expected Outcomes:	Report discussin gtradeffs and performance implications

Reference Material Links:	http:ertos/nicta.com/au

Project Title:	Video phone

Name of Supervisor:	Ihor Kuz

Email of Supervisor:	Ihor.kuz@nicta.com.au

Name of Joint/Co-Supervisor:	Nicholas FitzRoy-Dale

Email of Joint/Co-Supervisor:	nfd@cse.unsw.edu.au

School:	School of Computer Science and Engineering

For CSE and EET Projects:	NICTA Project

Faculty Research Area (Theme):	Embedded Systems and Communications

School Research Area:	Embedded, Real Time & Operating Systems

Applicable to other Engineering schools/disciplines:

Abstract:	A great way to demonstrate the flaws of a componentised operating system is to run a multimedia application on it (which may explain why there are so few of them). Two-way video conferencing is a particularly good demonstration because it uses a large number of operating system components (such as an IP stack and video, camera, and network drivers) and requires careful co-ordination between the scheduler and all threads in the system to manage the large amount of data flow. This project aims to produce a video conferencing application on a componentised operating system. It will involve some implementation work and design and testing of a complete system.

Research Environment:	The project will be conducted with the ERTOS group, one of the world's leading operating-systems research groups, in a stimulating environment that combines cutting-edge research with building systems that are deployed in real products.

Novelty and Contribution:	Besides contributing to the development of an embedded OS, this project will provide insights into the design issues and performance tradeoffs involved with componentising system services.

Expected Outcomes:	A video conferencing application on an embedded componentised operating system.

Reference Material Links:	ERTOS: http://www.ertos.nicta.com.au/ seL4: http://www.ertos.nicta.com.au/research/sel4/ L4: http://portal.ok-labs.com/ Componentised OS: http://www.ertos.nicta.com.au/research/camkes/ Course Prerequisites: Operating Systems. Advanced OS is highly recommended Skill Prerequisites: A good knowledge of C and system-level programming are necessary.

Project Title:	Web Server for L4-based Devices

Name of Supervisor:	Ihor Kuz

Email of Supervisor:	Ihor.kuz@nicta.com.au

Name of Joint/Co-Supervisor:	Nicholas FitzRoy-Dale

Email of Joint/Co-Supervisor:	nfd@cse.unsw.edu.au

School:	School of Computer Science and Engineering

For CSE and EET Projects:	NICTA Project

Faculty Research Area (Theme):	Embedded Systems and Communications

School Research Area:	Embedded, Real Time & Operating Systems

Applicable to other Engineering schools/disciplines:

Abstract:	Any computer system worth its salt must be able to run a web server these days. At ERTOS we are building a research OS based on a component architecture and a microkernel. We already have a (simple) network stack, but we still don't have a web-server! What this project will accomplish is to design and build a componentised web server and OS. But, it doesn't end there. The system must be fast. Therefore a significant amount of effort will also be spent analysing and optimising the resulting system.

Research Environment:	The project will be conducted with the ERTOS group, one of the world's leading operating-systems research groups, in a stimulating environment that combines cutting-edge research with building systems that are deployed in real products.

Novelty and Contribution:	Besides contributing to the further development of an embedded OS, this project will provide insights into the design issues and performance tradeoffs involved with componentising application and system services. The completed webserver will also be used to run parts of the ERTOS website.

Expected Outcomes:	A working web server integrated into our componentised OS. The insight into the design decisions and tradeoffs involved in designing such a subsystem will influence and further drive research into structuring modular, microkernel-based operating systems.

Reference Material Links:	ERTOS: http://www.ertos.nicta.com.au / seL4: http://www.ertos.nicta.com.au/research/sel4/ L4: http://portal.ok-labs.com http://www.l4hq.org/ http://www.l4ka.org/ Course Prerequisites: Operating Systems. Advanced OS is highly recommended Skill Prerequisites: A good knowledge of C and system-level programming are necessary.

Formal Methods

Project Title:	A Pragmatic, Formal Verification Framework for ARM Assembly Code in Isabelle/HOL

Name of Supervisor:	Gerwin Klein

Email of Supervisor:	kleing@cse.unsw.edu.au

Name of Joint/Co-Supervisor:	June Andronick

Email of Joint/Co-Supervisor:	june.andronick@nicta.com.au

School:	School of Computer Science and Engineering

For CSE and EET Projects:	NICTA Project

Faculty Research Area (Theme):	Embedded Systems and Communications

School Research Area:	Formal Methods

Applicable to other Engineering schools/disciplines:	Sciences – Maths, Physics, Chemistry

Abstract:	Modern operating system kernels such as seL4 are implemented mostly in C and, where performance or hardware access are critical, in assembly. The L4.verified project has created a formal verification framework for low-level C code. This project is about creating a similar framework for the practical formal verification of ARM assembly code in the interactive theorem prover Isabelle/HOL.

Research Environment:	You will work under guidance and with the support of international experts in the theorem proving and operating-systems fields, in a stimulating environment that combines cutting-edge research with building systems that are deployed in real products.

Novelty and Contribution:

Expected Outcomes:	An Assembly verification framework will strongly contribute to the state-of-the-art level of assurance that can be guaranteed of software. This project aims at being followed by an Honor thesis or PhD thesis that would provide a formal verification of seL4 assembly code.

Reference Material Links:	ERTOS: http://www.ertos.nicta.com.au l4.verified: http://www.ertos.nicta.com.au/research/l4.verified/ seL4: http://www.ertos.nicta.com.au/research/sel4/ Isabelle home page: http://isabelle.in.tum.de/

Project Title:	Formal Verification of Functional Correctness of User-Level Components in Microkernel Systems

Name of Supervisor:	Gerwin Klein

Email of Supervisor:	kleing@cse.unsw.edu.au

Name of Joint/Co-Supervisor:	June Andronick

Email of Joint/Co-Supervisor:	june.andronick@nicta.com.au

School:	School of Computer Science and Engineering

For CSE and EET Projects:	NICTA Project

Faculty Research Area (Theme):	Embedded Systems and Communications

School Research Area:	Formal Methods

Applicable to other Engineering schools/disciplines:	Sciences – Maths, Physics, Chemistry

Abstract:	The L4.verified project has produced a formal proof of correctness of the seL4 microkernel. This project fits in the aim of formally verifying, in the theorem prover Isabelle/HOL, the functional correctness of user-level components running on top of a microkernel. This requires turning our existing internal model of seL4 functionality into an external, user-level view of microkernel primitives and integrating this model with an existing verification framework for C programs.

Research Environment:	You will work under guidance and with the support of international experts in the theorem proving and operating-systems fields, in a stimulating environment that combines cutting-edge research with building systems that are deployed in real products.

Novelty and Contribution:

Expected Outcomes:	An initial formalisation on paper or in Isabelle/HOL of selected seL4 microkernel primitives from the user perspective. A formalisation of all seL4 kernel primitives from the internal kernel perspective exists already. The task is to describe what effect kernel calls have on user processes instead.

Reference Material Links:	ERTOS: http://www.ertos.nicta.com.au l4.verified: http://www.ertos.nicta.com.au/research/l4.verified/ seL4: http://www.ertos.nicta.com.au/research/sel4/ Isabelle home page: http://isabelle.in.tum.de/

Project Title:	Processor Modelling for Kernel Verification

Name of Supervisor:	Gerwin Klein

Email of Supervisor:	kleing@cse.unsw.edu.au

Name of Joint/Co-Supervisor:	David Cock

Email of Joint/Co-Supervisor:	davec@cse.unsw.edu.au

School:	School of Computer Science and Engineering

For CSE and EET Projects:	NICTA Project

Faculty Research Area (Theme):	Embedded Systems and Communications

School Research Area:	Formal Methods

Applicable to other Engineering schools/disciplines:	Sciences – Maths, Physics, Chemistry

Abstract:	The L4.verified project has produced a formal proof of correctness of the seL4 microkernel. The aim of this project is to extend our locally developed model of the ARM processor architecture to include privileged and system-management behaviour (processor modes, banked registers, memory management). This model underpins the bare-metal correctness proofs for the seL4 microkernel. The project may involve: simulator/HDL-compiler development, model validation on real hardware, formal model development in Isabelle/HOL.

Research Environment:	You will work under guidance and with the support of international experts in the theorem proving and operating-systems fields, in a stimulating environment that combines cutting-edge research with building systems that are deployed in real products.

Novelty and Contribution:

Expected Outcomes:	The exprected outcome of the project is an extended version of an existing ARM processor model for privileged mode.

Reference Material Links:	ERTOS: http://www.ertos.nicta.com.au l4.verified: http://www.ertos.nicta.com.au/research/l4.verified/ seL4: http://www.ertos.nicta.com.au/research/sel4/ Isabelle home page: http://isabelle.in.tum.de/

Project Title:	The Top 100 Theorems

Name of Supervisor:	Gerwin Klein

Email of Supervisor:	kleing@cse.unsw.edu.au

Name of Joint/Co-Supervisor:	Thomas Sewell

Email of Joint/Co-Supervisor:	thomas.sewell@nicta.com.au

School:	School of Computer Science and Engineering

For CSE and EET Projects:	NICTA Project

Faculty Research Area (Theme):	Embedded Systems and Communications

School Research Area:	Formal Methods

Applicable to other Engineering schools/disciplines:	Sciences – Maths, Physics, Chemistry

Abstract:	Of the top 100 most famous and interesting mathematical theorems in the world, 80 have now been verified in an automated proof assistant, but only 72 in one theorem prover. Will we be able to get the full 100? This summer project consists of picking a small number of the easier and smaller ones of these famous theorems and prove them in the interactive, industrial grade theorem prover Isabelle. Although this is a fun project, it has the more serious background of driving interactive theorem proving technology and to make this technology more accessible to not-yet-experts. You will learn state of the art theorem proving technology that is also applied to program verification.

Research Environment:	You will work under guidance and with the support of international experts in the theorem proving and operating-systems fields, in a stimulating environment that combines cutting-edge research with building systems that are deployed in real products.

Novelty and Contribution:

Expected Outcomes:	If successful, your work has an excellent chance of bein published in either the Archive of Formal Proof or the Isabelle theorem prover distribution directly.

Reference Material Links:	ERTOS: http://www.ertos.nicta.com.au Isabelle home page: http://isabelle.in.tum.de/ Archive of Formal Proofs: http://afp.sourceforge.net/ The Top 100 Theorems: http://www.cs.ru.nl/~freek/100/

Project Title:	User interface for a high-performance, retargetable architecture simulator

Name of Supervisor:	Gerwin Klein

Email of Supervisor:	kleing@cse.unsw.edu.au

Name of Joint/Co-Supervisor:	David Cock

Email of Joint/Co-Supervisor:	davec@cse.unsw.edu.au

School:	School of Computer Science and Engineering

For CSE and EET Projects:	NICTA Project

Faculty Research Area (Theme):	Embedded Systems and Communications

School Research Area:	Formal Methods

Applicable to other Engineering schools/disciplines:

Abstract:	The aim of this project is to develop a user interface for debugging system-level code in our locally-developed retargetable architectural simulator. You will evaluate the available options (Eclipse plugin, www.eclipse.org vs. Custom-built GUI), take into account usability requirements, and then deliver a frontend module to integrate with the existing (Python based) runtime framework.

Research Environment:	You will work under guidance and with the support of international experts in the theorem proving and operating-systems fields, in a stimulating environment that combines cutting-edge research with building systems that are deployed in real products.

Novelty and Contribution:

Expected Outcomes:	When successfully completed, it is intended that this frontend will be used in the low-level implementation of the next-generation secure microkernel seL4, as well as on other embedded systems work within NICTA, and be included in a publicly-released open-source version.

Reference Material Links:	ERTOS: http://www.ertos.nicta.com.au l4.verified: http://www.ertos.nicta.com.au/research/l4.verified/ seL4: http://www.ertos.nicta.com.au/research/sel4/ Isabelle home page: http://isabelle.in.tum.de/

Hardware Design, Computer Architectures, etc

Project Title:	A Memory Arbiter with Plug-and-Play Interface Protocols

Name of Supervisor:	Sri Parameswaran

Email of Supervisor:	sridevan@cse.unsw.edu.au

Name of Joint/Co-Supervisor:	Andhi Janapsatya

Email of Joint/Co-Supervisor:	andhij@cse.unsw.edu.au

School:	School of Computer Science and Engineering

For CSE and EET Projects:	School Project

Faculty Research Area (Theme):	Embedded Systems and Communications

School Research Area:	Hardware Design, Computer Architectures, etc

Applicable to other Engineering schools/disciplines:

Abstract:	The project involves the construction and testing of memory arbiters at the C/C++ level for use in embedded system simulators. The arbiter will handle different interface protocols, such as AMBA from ARM. These interface protocols should fit into the C/C++ models in a Plug-and-Play manner to allow fast and efficient modelling of SoC architectures. Students will understand the workings of several different memory interface protocols and gain experience in implementing these protocols in SoC similar to what is available in commercial products.

Research Environment:	The student will work with the aid of several post-docs, academic staff and PhD students.

Novelty and Contribution:	Existing simulators do not provide the necessary flexibility for use with different interface protocols.

Expected Outcomes:	A C/C++ based simulator of a memory arbiter and interface protocol models that can provide many benefits to future projects.

Reference Material Links:	http://www.micron.com/products/dram/

Project Title:	C/C++ Memory Controller Model

Name of Supervisor:	Sri Parameswaran

Email of Supervisor:	sridevan@cse.unsw.edu.au

Name of Joint/Co-Supervisor:	Andhi Janapsatya

Email of Joint/Co-Supervisor:	andhij@cse.unsw.edu.au

School:	School of Computer Science and Engineering

For CSE and EET Projects:	School Project

Faculty Research Area (Theme):	Embedded Systems and Communications

School Research Area:	Hardware Design, Computer Architectures, etc

Applicable to other Engineering schools/disciplines:

Abstract:	This project will develop an efficient simulator to control modern memory accesses. This involves the scheduling of requests into a control sequence for different memories based on their features. The student will analyse several algorithms for performing the scheduling to determine the sequences with the best performance.

Research Environment:	The student will work with the aid of several post-docs, academic staff and PhD students.

Novelty and Contribution:	Memory controller models do not exist at the software level. The completed design will aid the use of further projects that require optimisations of memory controllers.

Expected Outcomes:	A C/C++ based memory controller that can be utilised to provide many benefits to future projects.

Reference Material Links:	http://www.micron.com/products/dram/

Project Title:	Efficient DDR-RAM Simulator

Name of Supervisor:	Sri Parameswaran

Email of Supervisor:	sridevan@cse.unsw.edu.au

Name of Joint/Co-Supervisor:	Andhi Janapsatya

Email of Joint/Co-Supervisor:	andhij@cse.unsw.edu.au

School:	School of Computer Science and Engineering

For CSE and EET Projects:	School Project

Faculty Research Area (Theme):	Embedded Systems and Communications

School Research Area:	Hardware Design, Computer Architectures, etc

Applicable to other Engineering schools/disciplines:

Abstract:	The project involves the construction and testing of a C/C++ simulator that models the memory used in most modern embedded systems, DDR-RAM, including the newer DDR2 and DDR3 flavours. DDR-RAM utilises several interesting features that make it highly suitable for optimisation, such as multiple banks, burst access technology and aggressive power-saving features. The simulator will simultaneously calculate timing and power consumption of a given sequence of commands. The completed simulator will be used in a system of tools to optimise power consumption of multimedia applications. Throughout this project, the student will learn how modern memory systems operates and how to construct complex models of SoC devices.

Research Environment:	The student will work with the aid of several post-docs, academic staff and PhD students.

Novelty and Contribution:	Existing simulators do not provide the necessary accuracy and wealth of data that the tool created by this project requires. The tool will be used to optimise multimedia applications.

Expected Outcomes:	A C/C++ based simulator that can be utilised to provide many benefits to future projects.

Reference Material Links:	http://www.micron.com/products/dram/

Human Computer Interaction

Project Title:	Cognitive Load Analysis in Pen Interaction

Name of Supervisor:	Dr. Fang Chen

Email of Supervisor:	fang.chen@nicta.com.au

Name of Joint/Co-Supervisor:	Dr. Julien Epps

Email of Joint/Co-Supervisor:	(j.epps@unsw.edu.au

School:	School of Computer Science and Engineering

For CSE and EET Projects:	NICTA Project

Faculty Research Area (Theme):	Intelligent & Autonomous Systems

School Research Area:	Human Computer Interaction

Applicable to other Engineering schools/disciplines:

Abstract:	Cognitive load is variously described as the level of perceived mental effort associated with thinking and reasoning in a human. Tasks performed under stressful conditions, or requiring complex hand/eye coordination, can markedly increase cognitive load, and thus potentially interfere with other thought processes and consequently reduce overall task performance. Feasibility and efficiency of pen interaction is an interesting problem that has been extensively studied in the scenario of multimodal interfaces. However, little research exists investigating how pen gestures and handwriting inputs can be used to automatically estimate the cognitive load of a user in real-time. Little is known about what types of gesture/handwriting features and what types of underling modelling would be appropriate. The electronic map, as an example of information intensive media, provides many possibilities of pen-based interaction, where pen gestures could include clicking, drawing, tracking or much more complex traces. The incorporation of pen gestures and handwriting inputs for cognitive load measurement is particularly relevant to the context of emergency services operators, who need to deal with many map applications and frequently write down notes about incidents as memory aids. If a user’s level of experiencing cognitive load can be assessed in real time via pen interaction, an intelligent human-machine interaction system could potentially change its reactions based on user’s responses. This can also lead to other real-life applications, e.g. in distance learning, and future mobile phone applications.

Research Environment:	The selected student will work with a team of senior researchers in the field of signal processing and human-computer interaction, at the NICTA premise inside the Australian Technology Park.

Novelty and Contribution:	Cognitive load measurement via pen interaction is a brand new and promising field of research. This work will have big impact on the research agenda of objective cognitive load mesaurement. It will enable researchers and developers to build intelligent human-machine systems that are adaptive to a user’s cognitive load.

Expected Outcomes:	The main research is to design experiments to capture potential pen-based features for cognitive load measurement and to analysis the data set to validate some of the features. You will work with a team of researchers on developing experiments on capturing pen gesture changes as cognitive load increases. Some literature review will be required to learn the basics of pen gesture production. Signal processing and statistical analysis will also be required to extract features and investigate correlations between cognitive load and pen interaction behaviour. You may have the opportunity to publish the results if they are of good quality.

Reference Material Links:	Contact supervisor

Project Title:	Galvanic Skin Response and Cognitive Load: Exploring things that make us break out in a sweat.

Name of Supervisor:	Natalie Ruiz

Email of Supervisor:	natalie.ruiz@nicta.com.au

Name of Joint/Co-Supervisor:	Fang Chen

Email of Joint/Co-Supervisor:	fang.chen@nicta.com.au

School:	School of Computer Science and Engineering

For CSE and EET Projects:	NICTA Project

Faculty Research Area (Theme):	Intelligent & Autonomous Systems

School Research Area:	Human Computer Interaction

Applicable to other Engineering schools/disciplines:

Abstract:	Cognitive load measurement is gaining increasing importance in the design and evaluation of all human/computer related products and processes, both traditional and technology based. Although it is well understood as a theoretical construct, the measurement of cognitive load mainly relies on methods that are either indirect, subjective, or both. Integrating aspects of cognitive load measurement, and how to detect users’ cognitive load levels by using non-intrusive methods under their real working environment are very interesting problems to solve, and can often be critical issues. So far the bulk of the work related to cognitive load measurement has been centered on speech signal processing based methods. This is because speech can be collected and processed in real-time, in a non-intrusive way. This project aims to examine the sensitivity of galvanic skin response to different levels of cognitive load. Galvanic skin response measures the conductance of your skin, or the way sweat glands are activated when people expend mental effort on puzzles or other complex tasks. It is expected that the successful applicant will collect and analyse GSR datasets, designing suitable tasks, collecting GSR readings under no load and various levels of high load, and use various commercial (Thought Technology) or in-house developed tools and techniques (or even develop their own tools as needed) to test the sensitivity of GSR to people under varying load levels. Alternatively, previously collected datasets may be provided to the student for analysis.

Research Environment:	You will work with a team of HCI researchers on the design of a study to collect GSR. Some literature review will be required to learn the basics of skin conductance technologies. Some data labelling will also be required. Results may be publishable, giving you the opportunity to take part in the entire research loop, from innovation to publication.

Novelty and Contribution:	The student will contribute to the research agenda by assessing the sensitivity and diagnostic value of GSR as a robust index of cognitive load, defining appropriate thresholds and context combinations.

Expected Outcomes:	The expected outcomes include: • Literature Scan of GSR • Analysis of GSR data for a previously collected dataset (Driving study) • Design and execution of a new user study, designed specifically to collect GSR readings under varying levels of cognitive load, perhaps even longitudinally. • Possible development of tools to label, segment or extract aforementioned GSR features

Reference Material Links:	Contact supervisor

Project Title:	Harder than the eye can see – Eye-gaze patterns under High Cognitive Load

Name of Supervisor:	Natalie Ruiz

Email of Supervisor:	natalie.ruiz@nicta.com.au

Name of Joint/Co-Supervisor:	Fang Chen

Email of Joint/Co-Supervisor:	fang.chen@nicta.com.au

School:	School of Computer Science and Engineering

For CSE and EET Projects:	NICTA Project

Faculty Research Area (Theme):	Intelligent & Autonomous Systems

School Research Area:	Human Computer Interaction

Applicable to other Engineering schools/disciplines:

Abstract:	Cognitive load measurement is gaining increasing importance in the design and evaluation of all human/computer related products and processes, both traditional and technology based. Although it is well understood as a theoretical construct, the measurement of cognitive load mainly relies on methods that are either indirect, subjective, or both. Integrating aspects of cognitive load measurement, and how to detect users’ cognitive load levels by using non-intrusive methods under their real working environment are very interesting problems to solve, and can often be critical issues. So far the bulk of the work related to cognitive load measurement has been centered on speech signal processing based methods. This is because speech can be collected and processed in real-time, in a non-intrusive way. This project aims to examine semantic eye-gaze patterns (what people look at, when and why) under different levels of cognitive load in a video-based cognition task. It is expected that the successful applicant will analyse eye-gaze data using in-house developed tools and techniques (or even develop their own tools as needed) to extract visual search patterns related to high cognitive load. In parallel, machine learning algorithms could be applied to the dataset to identify differences between different load levels.

Research Environment:	You will work with a team of HCI researchers to develop a series of hypotheses related to the effect of cognitive load on task specific visual search strategies. Some literature review will be required to learn the basics of eye-gaze research, measurable features such as fixations and saccades, as well as general visual search strategies, attention and fatigue related literature. Some data labelling may also be required. Results may be publishable, giving you the opportunity to take part in the entire research loop, from innovation to publication.

Novelty and Contribution:	The student will contribute to the research agenda by assessing the feasibility of using a new modality to a suite of cognitive load measurement indices.

Expected Outcomes:	The expected outcomes include: • Literature Scan of Eye-gaze research • Analysis of Eye-Gaze data for a previously collected dataset using available tools and techniques • Apply machine learning algorithms to eye-gaze data to classify and identify visual strategies depending on load levels. • Possible development of tools to label, segment or extract aforementioned eye-gaze features

Reference Material Links:	Requirements for NICTA Scholarship: Excellent video processing skills, programming skills in java or C, interest in HCI or cognitive science, good communication skills, and an inquisitive and enthusiastic attitude.

Project Title:	Loaded gestures: Automating analysis of Manual Gesture for Cognitive Load Detection

Name of Supervisor:	Natalie Ruiz

Email of Supervisor:	natalie.ruiz@nicta.com.au

Name of Joint/Co-Supervisor:	Fang Chen

Email of Joint/Co-Supervisor:	fang.chen@nicta.com.au

School:	School of Computer Science and Engineering

For CSE and EET Projects:	NICTA Project

Faculty Research Area (Theme):	Intelligent & Autonomous Systems

School Research Area:	Human Computer Interaction

Applicable to other Engineering schools/disciplines:

Abstract:	Cognitive load measurement is gaining increasing importance in the design and evaluation of all human/computer related products and processes, both traditional and technology based. Although it is well understood as a theoretical construct, the measurement of cognitive load mainly relies on methods that are either indirect, subjective, or both. Integrating aspects of cognitive load measurement, and how to detect users’ cognitive load levels by using non-intrusive methods under their real working environment are very interesting problems to solve, and can often be critical issues. So far the bulk of the work related to cognitive load measurement has been centered on speech signal processing based methods. This is because speech can be collected and processed in real-time, in a non-intrusive way. This project aims to examine the patterns of manual gesture during both interaction and conversation, to complement speech based measures of load. Specifically, we would like to explore whether general gesture features such as gesture type, amplitude or frequency can tell us anything about whether a person is cognitively loaded or not. It is expected that the successful applicant will analyse gesture datasets from a variety of contexts (netball testing, map-based interaction, and others) using various in-house developed tools and techniques (e.g. skin colour detection, hand tracking); or even develop their own tools as needed.

Research Environment:	You will work with a team of HCI researchers on developing hypotheses about the way gesture changes as load increases. Some literature review will be required to learn the basics of gesture production. Some data labelling will also be required. Results may be publishable, giving you the opportunity to take part in the entire research loop, from innovation to publication.

Novelty and Contribution:	The student will contribute to the research agenda by proposing and testing potential gesture-based features for cognitive load measurement.

Expected Outcomes:	The expected outcomes include: • Development of hypotheses based on cognitive load and gesture • Literature Scan of Gesture production • Analysis of gesture-based features • Possible development of tools to label, segment or extract aforementioned features

Reference Material Links:	Excellent video or image processing skills, programming skills in java or C, interest in HCI or cognitive science, good communication skills, and an inquisitive and enthusiastic attitude.

Project Title:	Phonemote – Turning Mobile Phones into Wii-like Game Remote Controllers

Name of Supervisor:	Arcot Sowmya

Email of Supervisor:	sowmya@cse.unsw.edu.au

Name of Joint/Co-Supervisor:

Email of Joint/Co-Supervisor:

School:	School of Computer Science and Engineering

For CSE and EET Projects:	School Project

Faculty Research Area (Theme):	Intelligent & Autonomous Systems

School Research Area:	Human Computer Interaction

Applicable to other Engineering schools/disciplines:	Biomedical Engineering Electrical Engineering & Telecommunications

Abstract:	Human-computer interaction (HCI) is concerned with the design, evaluation and implementation of interactive computing systems for human use and the study of the major phenomena surrounding them. The goal of this project is to develop new perceptual interfaces for human-computer-interaction based on visual input captured by mobile phone cameras, and to investigate how such interfaces can complement or replace traditional interfaces based on keyboards, mics, remote control devices, data gloves or Wii Remote

Research Environment:	The research team at CSE contains a senior academic and about 7 PhD students / p ostdocs. The team has already developed a number of systems for image segmentation and recognition and dynamic models for video tracking. These provide a working platform to extend and apply to the HCI interfaces propsoed in this project.

Novelty and Contribution:	This research may result in new approaches of human computer interaction systems that have a large range of applications from computer input to entertainment equipment control.

Expected Outcomes:	Outcomes of this research include software packages and a short technical report documenting the research. The work may be publishable.

Reference Material Links:	contact supervisor

Project Title:	Smart Room Interaction Environment

Name of Supervisor:	Arcot Sowmya

Email of Supervisor:	sowmya@cse.unsw.edu.au

Name of Joint/Co-Supervisor:

Email of Joint/Co-Supervisor:

School:	School of Computer Science and Engineering

For CSE and EET Projects:	School Project

Faculty Research Area (Theme):	Immersive Systems and Virtual Reality

School Research Area:	Human Computer Interaction

Applicable to other Engineering schools/disciplines:	Biomedical Engineering Electrical Engineering & Telecommunications

Abstract:	Smart rooms provide a novel user interface to a physical or virtual environment. In this summer research project, you will be implementing an interaction mechanism for a Smart Room meeting environment, using computer vision techniques. The system will track users within the smart room using 4 static colour cameras and a pan-tilt-zoom camera as sensors. The system will then provide a method for users to interact with the room itself. Examples of such interactions include pointing to a light switch to turn it on or performing a certain gesture to turn on the projector.

Research Environment:	The research team at CSE contains a senior academic and about 7 PhD students / postdocs. The team has already developed many systems that perform image segmentation and recognition and build dynamic models for video tracking. These provide a working platform to extend and apply the techniques to smart room interactions. A prototype smart room, with PTZ cameras on overhead tracks, and servers that control them, already exists in CSE, and will be used in the project.

Novelty and Contribution:	This research may result in new Smart Room technology, which provides unique interaction mechanisms within the rooms.

Expected Outcomes:	Outcomes of this project include an interaction application for the smart room within CSE.

Reference Material Links:	contact supervisor

Image Processing

Project Title:	An improved face detector

Name of Supervisor:	Getian Ye

Email of Supervisor:	Getian.Ye@nicta.com.au

Name of Joint/Co-Supervisor:	Yang Wang, Bang Zhang

Email of Joint/Co-Supervisor:	Yang.Wang@nicta.com.au, Bang.Zhang@nicta.com.au

School:	School of Computer Science and Engineering

For CSE and EET Projects:	NICTA Project

Faculty Research Area (Theme):	Intelligent & Autonomous Systems

School Research Area:	Image Processing

Applicable to other Engineering schools/disciplines:	Electrical Engineering & Telecommunications

Abstract:	Detecting faces in an image is a fundamental task for many promising applications, e.g., visual surveillance, content-based image retrieval, and robotics. Viola and Jones’ work is regarded as the latest milestone in the research of face detection. They developed an accurate frontal face detection system that employs Haar features and AdaBoost learning process. Following their work, many improved solutions have been proposed. The aim of this project is to develop and implement an improved face detector that is recently proposed by researchers at NICTA. The successful candidate will improve skill in problem solving and gain basic knowledge of image processing and the state-of-the-art face detection methods.

Research Environment:	The successful candidate will be working with researchers and postgraduate students at Kensington Laboratory, National ICT Australia.

Novelty and Contribution:	The novelty lies in improving face detection using discriminative visual features. Part of the outcome of this project will contribute to NICTA’s Strategic Project – Smart Transport and Road (STaR).

Expected Outcomes:	Software implementation, experiments, and demo.

Reference Material Links:	http://en.wikipedia.org/wiki/Face_detection

Project Title:	Computer simulation of road traffic geometry

Name of Supervisor:	Yang Wang

Email of Supervisor:	yang.wang@nicta.com.au

Name of Joint/Co-Supervisor:	Getian Ye

Email of Joint/Co-Supervisor:	getian.ye@nicta.com.au

School:	School of Computer Science and Engineering

For CSE and EET Projects:	NICTA Project

Faculty Research Area (Theme):	Intelligent & Autonomous Systems

School Research Area:	Image Processing

Applicable to other Engineering schools/disciplines:	Electrical Engineering & Telecommunications

Abstract:	In camera based visual surveillance/traffic monitoring, objects within a 3-D scene together with their spatial relationships are projected onto a sequence of 2-D camera images. This work is to simulate the geometry relationship between the 3-D real scene and the 2-D camera image for road traffic monitoring. Clearly, both the structure of the 3-D scene and the projection from 3-D to 2-D (or image formation) will influence the result of geometry analysis. The geometry relationship can be studied with simplifications through computer simulation.

Research Environment:	The successful candidate will work with researchers and postgraduate students at NICTA’s Kensington Laboratory.

Novelty and Contribution:	The project will deal with the geometry analysis for the cases such as curved road and camera shake. The research outcome will contribute to NICTA’s Smart Transport and Roads (STaR) project.

Expected Outcomes:	Software implementation and demo.

Reference Material Links:	http://www.nicta.com.au/research/projects/smart_transport_and_roads/star_projects

Project Title:	Image enhancement with a few high quality images

Name of Supervisor:	Getian Ye

Email of Supervisor:	Getian.Ye@nicta.com.au

Name of Joint/Co-Supervisor:	Yang Wang

Email of Joint/Co-Supervisor:	Yang.Wang@nicta.com.au

School:	School of Computer Science and Engineering

For CSE and EET Projects:	NICTA Project

Faculty Research Area (Theme):	Intelligent & Autonomous Systems

School Research Area:	Image Processing

Applicable to other Engineering schools/disciplines:	Biomedical Engineering Electrical Engineering & Telecommunications

Abstract:	The quality of a digital image is usually affected by many factors such as the size of camera sensor, noises, and environmental conditions. For example, the sensor of a webcam is usually small and noisy so that the image quality is low. When a camera is used in an uncontrolled environment, different artefacts like dust and blurring may be introduced in the image. Improving image quality can not only give the viewer more pleasing pictures but can also offer more details that may be critical in different kinds of applications. This project aims to develop a novel technique for image enhancement using a few high quality images, which are independent of the image to be improved. The nature of the work is mainly software development. The successful candidate will improve skill in problem solving and gain basic knowledge of image processing.

Research Environment:	The successful candidate will be working with researchers and postgraduate students at Kensington Laboratory, National ICT Australia.

Novelty and Contribution:	The novelty lies in improving image quality with unrelated high quality images. Part of the outcome of this project will contribute to NICTA’s Strategic Project – Smart Transport and Road (STaR).

Expected Outcomes:	Software implementation, experiments, and demo.

Reference Material Links:

Project Title:	Learn to detect vehicles using low-level features

Name of Supervisor:	Getian Ye

Email of Supervisor:	Getian.Ye@nicta.com.au

Name of Joint/Co-Supervisor:	Yang Wang

Email of Joint/Co-Supervisor:	Yang.Wang@nicta.com.au

School:	School of Computer Science and Engineering

For CSE and EET Projects:	NICTA Project

Faculty Research Area (Theme):	Intelligent & Autonomous Systems

School Research Area:	Image Processing

Applicable to other Engineering schools/disciplines:

Abstract:	Vehicle detection is the most important image understanding task in video-based traffic monitoring systems. Effective vehicle detection is helpful for vehicle tracking, classification, and event analysis. Moreover, it improves traffic control and reduces congestion. This project aims to develop an approach to vehicle detection using low-level image features. The nature of the work is mainly software development. The successful candidate will improve skill in problem solving and gain basic knowledge of image processing, pattern recognition, and machine learning.

Research Environment:	The candidate will be working with researchers and postgraduate students at Kensington Laboratory, National ICT Australia.

Novelty and Contribution:	The novelty lies in improving vehicle detection using low-level image features and machine learning techniques.

Expected Outcomes:	Part of the outcome of this project will contribute to NICTA’s Strategic Project – Smart Transport and Road (STaR).

Reference Material Links:	http://www.nicta.com.au/research/projects/smart_transport_and_roads

Project Title:	Object detection using stereo camera

Name of Supervisor:	Yang Wang

Email of Supervisor:	yang.wang@nicta.com.au

Name of Joint/Co-Supervisor:	Getian Ye

Email of Joint/Co-Supervisor:	getian.ye@nicta.com.au

School:	School of Computer Science and Engineering

For CSE and EET Projects:	NICTA Project

Faculty Research Area (Theme):	Intelligent & Autonomous Systems

School Research Area:	Image Processing

Applicable to other Engineering schools/disciplines:	Electrical Engineering & Telecommunications

Abstract:	Object detection is an important image understanding task with many potential applications. For example, in camera based visual surveillance/traffic monitoring systems, faces/vehicles are usually required to be robustly detected under various circumstances. Stereo camera can simultaneously capture a pair of images of the same scene and provide information for distinguishing the objects at different ranges in the scene. This project is to develop and implement an object detection algorithm using stereo camera.

Research Environment:	The successful candidate will work with researchers and postgraduate students at NICTA’s Kensington Laboratory.

Novelty and Contribution:	Stereo camera helps to improve the robustness of object detection under varying illumination and weather conditions. The research outcome will contribute to NICTA’s Smart Transport and Roads (STaR) project.

Expected Outcomes:	Software implementation and demo.

Reference Material Links:	http://www.videredesign.com/vision/svs_images.htm

Project Title:	Pedestrian Detection Using a Cascade of Boosted Classifiers

Name of Supervisor:	Jian Zhang

Email of Supervisor:	jzhang@cse.unsw.edu.au

Name of Joint/Co-Supervisor:

Email of Joint/Co-Supervisor:

School:	School of Computer Science and Engineering

For CSE and EET Projects:	NICTA Project

Faculty Research Area (Theme):	Intelligent & Autonomous Systems

School Research Area:	Image Processing

Applicable to other Engineering schools/disciplines:	Electrical Engineering & Telecommunications

Abstract:	This project aims provide an implementation, verification and evaluation of a pedestrian detection framework proposed by Viola-Jones which is fast and robust under varying conditions by incorporating both motion and appearance in training a cascade classifiers. It is hoped that this project will serve as a foundation for other applications such as vehicle and cyclist detection; and incorporated with other image processing techniques such as tracking which would improve robustness and computational efficiency. The nature of this work will focus on software implementation based on state-of-art international conference papers.

Research Environment:	The novelty lies on the new approach for image categorization based on feature extraction. You will access the current state-of-art image/video processing in-house-made tools. Your work will contribute to NICTA’s project – Video analysis and content management for surveillance (VACMS).

Novelty and Contribution:	The project will be carried out in multimedia & video communication research group in National ICT Australia (NICTA) laboratories at Kensington. You will work directly with many talent researchers in a high standard image/video group and will be given clear guidance for work that could link to your future career development

Expected Outcomes:	Algorithm software implementation and some test results.

Reference Material Links:	Paul Viola, Michael J. Jones, Daniel Snow. "Detecting Pedestrians using Patterns of Motion and Appearance", ICCV 2003. November 27 http://www.merl.com/papers/docs/TR2003-90.pdf

Miscellaneous

Project Title:	Software for Advanced Patent Analysis

Name of Supervisor:	Vladimir Tosic

Email of Supervisor:	vtosic@cse.unsw.edu.au

Name of Joint/Co-Supervisor:	Mark Staples

Email of Joint/Co-Supervisor:	Mark.Staples@nicta.com.au

School:	School of Computer Science and Engineering

For CSE and EET Projects:	NICTA Project

Faculty Research Area (Theme):	Management

School Research Area:	Miscellaneous

Applicable to other Engineering schools/disciplines:	Biomedical Engineering Chemical Sciences and Engineering Civil & Environmental Engineering Electrical Engineering & Telecommunications Mechanical & Manufacturing Engineering Mining Engineering Petroleum Engineering Photovoltaic and Renewable Energy Engineering Surveying & Spatial Information Systems Sciences – Maths, Physics, Chemistry

Abstract:	To protect its intellectual property, it is often necessary for a company to patent its inventions. Patents are legally enforceable rights for exclusive commercial exploitation of inventions. Before patenting, patent search and analysis can uncover important facts relevant for strategic decisions about company’s intellectual property and research and development activities in general. Various software tools support patent search and analysis, from relatively simple free tools and Web sites to more powerful commercial products (e.g., for determining and visualizing various dependencies). In this research project, students will help develop novel software for advanced patent analysis, based on a new patent analysis methodology. The methodology is currently supported by software that manages patent information in Excel and uses macros for processing and visualization of patent analyses. The first aspect of this project is to support querying and analysis of patent information stored in a relational database. The second aspect of the project is to implement additional advanced patent analysis procedures. The third aspect of this project involves search and analysis of a number of real patents, determining their characteristics, and storing and managing this information using the developed software tool, to evaluate the tool’s correctness and usefulness.

Research Environment:	The students will work closely with researchers at NICTA (http://www.nicta.com.au) in a friendly mixed-gender and multicultural environment comprised of senior researchers and postgraduate students.

Novelty and Contribution:	The main novelty is the support for a unique and new patent analysis methodology. Since some aspects of the new patent procedures have not been implemented previously in other systems, non-trivial research questions (e.g., how to categorize patents in terms of relevance for company’s business strategy) will have to be considered. These patent analysis procedures will enable better decision making about a company’s patent portfolio. Another contribution is the testing process, which will result in conclusions about real patents from one market area (e.g., implant systems, business-driven IT systems management, or another area of mutual interest).

Expected Outcomes:	- Architecture of a software system that stores patent information, processes this information (e.g., to determine various dependencies), and visualizes results. - Detailed design of modules of this software architecture. - Design of database for storing patent information. - Original patent analysis procedures, which query and process the stored patent information. - Implementation of the above-mentioned designs. - Design and implementation of a simple (possibly Web) interface into the system. - Population of the database with patent information for a number of real patents from the same scientific area. - Experiments evaluating correctness and usefulness of the implemented software.

Reference Material Links:	- http://en.wikipedia.org/wiki/Patent - http://www.ipaustralia.gov.au/patents/what_index.shtml - http://www.google.com/patents - http://www.patentlawlinks.com/patsearc.htm - http://www.infovis.net/printMag.php?lang=2&num=167 - D. Hunt, L. Nguyen, M. Rodgers (Eds.) “Patent Searching: Tools & Techniques”, Wiley, 2007 - J.L. Davis, S.S. Harrison “Edison in the Boardroom: How Leading Companies Realize Value from Their Intellectual Assets”, Wiley, 2001 - Course COMP9311 “Database Systems” (http://www.cse.unsw.edu.au/~cs9311) - http://www.edumax.com/database-basics-chapter-2-the-er-model-and-database-design.html - http://www.w3schools.com/SQl/default.asp - Course COMP9321 “Web Applications Engineering” (http://www.cse.unsw.edu.au/~cs9321) - For further information, email Dr. Vladimir Tosic (‘vtosic’ at the CSE e-mail system) with Subject line “UNSW Summer Scholars”.

Multimedia & Visual Communication

Project Title:	A Real-time and Robust Object Tracking System for the Video Surveillance.

Name of Supervisor:	Jian Zhang

Email of Supervisor:	jian.zhang@nicta.com.au

Name of Joint/Co-Supervisor:	Sijun Lu

Email of Joint/Co-Supervisor:	sijun.lu@nicta.com.au

School:	School of Computer Science and Engineering

For CSE and EET Projects:	NICTA Project

Faculty Research Area (Theme):	Intelligent & Autonomous Systems

School Research Area:	Multimedia & Visual Communication

Applicable to other Engineering schools/disciplines:	Electrical Engineering & Telecommunications

Abstract:	The advance of fast and affordable computing device has provided the power to track moving objects from a video camera in real-time. Many algorithms have been developed for object tracking in the past, using various image/video features, such edge, color, spatial-temporal information, or using Kalman filtering and hypothesis testing techniques. The nature of this work is to implement the existing object tracking techniques and choose one suitable for the video surveillance domain and the real-time purpose. Research challenge of this task is to achieve robust tracking under the disturbing environment such as occlusion and lighting changes. The knowledge you learned from the computer and engineering courses, such as signal processing, graphic theory and computer vision will gain plenty practice from this project. Your programming skills will also be developed as you will develop a demo in software. This project is an absolute opportunity for a ToR student to test his/her research capabilities for the PhD study.

Research Environment:	The project will be carried out in multimedia & video communication research group in National ICT Australia (NICTA) laboratories at Kensintong. You will work directly with many talent researchers in a high standard image/video group and will be given clear guidance for work that could link to your future career development.

Novelty and Contribution:	The novelty lies on the new approach to tracking objects without implementing background subtraction. You will access the current state-of-art image/video processing in-house-made tools. Your work will contribute to NICTA’s Strategic Project – Smart Transport and Road (STaR).

Expected Outcomes:	Algorithm software implementation and some test results.

Reference Material Links:	[1] T. Kanade, R. Collins, A. Lipton, P. Burt, and L. Wixson, "Advances in Cooperative Multi-Sensor Video Surveillance",Darpa Image Understanding Workshop, Morgan Kaufmann, November, 1998, pp. 3-24 [2]R. Collins, A. Lipton, H. Fujiyoshi, and T. Kanade, "Algorithms for cooperative multisensor surveillance", Proceedings of the IEEE, Vol. 89, No. 10, October, 2001, pp. 1456 - 1477. [3] Tao Xiong and Christian Debrunner, “Stochastic Car Tracking With Line- and Color-Based Features”, IEEE Transactions on Intelligent Transportation Systems, VOL. 5, NO. 4, December 2004 [4] Yue Zhou and Hai Tao, “A Background Layer Model for Object Tracking through Occlusion”, ICCV 2003

Project Title:	Appearance-based re-identification of objects in video

Name of Supervisor:	Jian Zhang

Email of Supervisor:	jian.zhang@nicta.com.au

Name of Joint/Co-Supervisor:

Email of Joint/Co-Supervisor:

School:	School of Computer Science and Engineering

For CSE and EET Projects:	NICTA Project

Faculty Research Area (Theme):	Intelligent & Autonomous Systems

School Research Area:	Multimedia & Visual Communication

Applicable to other Engineering schools/disciplines:	Electrical Engineering & Telecommunications

Abstract:	This research is of prime importance in many video content management applications including video surveillance, image/video retrieval and event detection in video. The premise of this task is that given an object of the video how can we re-identify the same object based on the appearance. This is a very challenge research in the video content analysis. Student will develop their academic research skills in the area of image processing/computer vision, pattern recognition and machine learning. The theme of the research project is called “use inspired research”. The PhD project will be potentially connected to the industrial research groups for future collaborations.

Research Environment:	The project will be carried out in multimedia & video communication research group in National ICT Australia (NICTA) laboratories at Kensington. You will work directly with many talent researchers in a high standard image/video group and will be given clear guidance for work that could link to your future career development.

Novelty and Contribution:	The novelty lies on the new approach for key feature detection, objects classification and multi-camera data fusion. You will access the current state-of-art image/video processing in-house-made tools. Your work will contribute to NICTA’s project in video analysis and event detection.

Expected Outcomes:	Algorithm software implementation and some test results.

Reference Material Links:	[1] "Background-Subtraction in Thermal Imagery Using Contour Saliency" J. Davis and V. Sharma International Journal of Computer Vision, Vol. 71, No. 2, 2007, pp. 161-181. (online version June 2006) [2] "A Two-Stage Template Approach to Person Detection in Thermal Imagery" J. Davis and M. Keck Workshop on Applications of Computer Vision, Breckenridge, Colorado, January 5-7, 2005. [3] Sakrapee (Paul) Paisitkriangkra, Chunhua Shen and Jian Zhang, “Fast Pedestrian Detection Using a Cascade of Boosted Covariance Features”, IEEE Trans. on Circuits and Systems for Video Technology, 18(8), pages 1140-1151, August 2008

Project Title:	Objects Classification and Event Detection in Surveillance Video

Name of Supervisor:	Jian Zhang

Email of Supervisor:	jian.zhang@nicta.com.au

Name of Joint/Co-Supervisor:	Sijun Lu

Email of Joint/Co-Supervisor:	sijun.lu@nicta.com.au

School:	School of Computer Science and Engineering

For CSE and EET Projects:	NICTA Project

Faculty Research Area (Theme):	Intelligent & Autonomous Systems

School Research Area:	Multimedia & Visual Communication

Applicable to other Engineering schools/disciplines:	Electrical Engineering & Telecommunications

Abstract:	Object classification is a further step to object detection and tracking in understanding of video. Without object classification, we can only know something is moving in the video sequences, but don’t know what it is. With the information from object tracking and classification, it may become possible to understand the activities and events happening in the video sequences. The nature of the work is software implementation to investigate the state-of-art object classification technologies for categorizing moving vehicles or human detected from the surveillance videos. The knowledge you learned from the courses of computer vision, neural network, pattern recognition and machine learning can all find their utilization in this project. Your programming skills will also be developed as you will develop a demo in software. This project is an absolute opportunity for a ToR student to test his/her research capabilities for the PhD study.

Research Environment:	The project will be carried out in multimedia & video communication research group in National ICT Australia (NICTA) laboratories at Kensington. You will work directly with many talent researchers in a high standard image/video group and will be given clear guidance for work that could link to your future career development.

Novelty and Contribution:	The novelty lies on the new approach for objects classification and statistical feature clustering. You will access the current state-of-art image/video processing in-house-made tools. Your work will contribute to NICTA’s Strategic Project – Smart Transport and Road (STaR).

Expected Outcomes:	Algorithm software implementation and some test results.

Reference Material Links:	[1] T. Kanade, R. Collins, A. Lipton, P. Burt, and L. Wixson, "Advances in Cooperative Multi-Sensor Video Surveillance",Darpa Image Understanding Workshop, Morgan Kaufmann, November, 1998, pp. 3-24 [2]R. Collins, A. Lipton, H. Fujiyoshi, and T. Kanade, "Algorithms for cooperative multisensor surveillance", Proceedings of the IEEE, Vol. 89, No. 10, October, 2001, pp. 1456 - 1477. [3] Wei Niu, Jiao Long, Dan Han and Yuan-Fang Wang, “Human Activity Detection and Recognition for Video Surveillance”, ICME 2004 [4] Surendra Gupte, Osama Masoud, Robert F. K. Martin, and Nikolaos P. Papanikolopoulos, “Detection and Classification of Vehicles”, IEEE Transactions on Intelligent Transportation Systems, VOL. 3, NO. 1, December 2002

Networks, Sensor Networks, etc

Project Title:	A Novel Privacy Preserving Technique for e-Commerce

Name of Supervisor:	Salil Kanhere

Email of Supervisor:	salilk@cse.unsw.edu.au

Name of Joint/Co-Supervisor:

Email of Joint/Co-Supervisor:

School:	School of Computer Science and Engineering

For CSE and EET Projects:	School Project

Faculty Research Area (Theme):	Embedded Systems and Communications

School Research Area:	Networks, Sensor Networks, etc

Applicable to other Engineering schools/disciplines:

Abstract:	Mobile e-commerce is expected to generate significant interest and revenue in the next few years. However, in order for mobile e-commerce to take off, it is imperative that the privacy of users is protected. In particular, identity theft is a serious problem, as is evident by the significant economic losses due to stolen private information (such as credit card and bank account details). Lockstep Technologies has recently developed a novel solution for dealing with identity theft called Stepwise, which uses specially formatted digital certificates. The ideas use basic Public Key Cryptography techniques to engineer a simple solution, which does not require users to reveal any personal information to the merchants. This technology has been deployed successfully on smart cards [1]. In this project we seek to migrate these ideas and implement the Stepwise algorithms on a mobile device.

Research Environment:	The student will work in conjunction with researchers at the Networks Research Lab. Student will also receive support from system architects at Lockstep Technologies, a sydney-based firm specialising in privacy and security.

Novelty and Contribution:	The proposed implementation will be the first implementation of the proposed algorithms on a mobile phone.

Expected Outcomes:	The project will involve developing two key components: (1) software on the mobile client, which implements the digital signature algorithms and (2) a back-end server, which implements the merchant server for verifying the signature.

Reference Material Links:	Reference Material: [1] Lockstep Technology Notes: http://www.lockstep.com.au/technologies/technology_notes [2] Java Security and Trust API: http://java.sun.com/j2me/docs/satsa-dg/pki.html http://developers.sun.com/mobility/apis/articles/satsa2/

Project Title:	Analyzing data from EnergyAustralia's WiMAX deployment

Name of Supervisor:	Athanassios Boulis

Email of Supervisor:	athanassios.boulis@nicta.com.au

Name of Joint/Co-Supervisor:

Email of Joint/Co-Supervisor:

School:	School of Computer Science and Engineering

For CSE and EET Projects:	NICTA Project

Faculty Research Area (Theme):	Embedded Systems and Communications

School Research Area:	Networks, Sensor Networks, etc

Applicable to other Engineering schools/disciplines:

Abstract:	EnergyAustralia has deployed a testbed of 5 WiMAX basestations in Newcastle as part of their smart grid initiative and they already have 20,000 measurements around the area. They need to better understand the deployment and answer question like "what is the expected coverage if we move to other cities?" "how can I change my deployment to achieve better results with comparable cost?". You will first process the data in various ways so you can cluster them and provide meta-information not currently available such as 1)basestation association, 2) traffic conditions 3)terrain type. You will then analyze the richer data based on some techniques proposed by your supervisor and be able to infer high level trends.

Research Environment:	You will be closely guided and supported by your academic supervisor having daily interactions. You will also have the opportunity to work along with other junior researchers in the research group. You will receive full support in terms or resources (computers, tools, IT services), enjoy the relaxed workplace, and take advantage of the seminars and workshops organised at NICTA to expand your academic and industry knowledge.

Novelty and Contribution:	Working with large data sets gathered from a real deployment

Expected Outcomes:	Meta-tagging the data thus providing more information to analysis. Results from some known analysis techniques.

Reference Material Links:	Contact supervisor

Project Title:	Enhancing models of a Wireless Sensor Network Simulator

Name of Supervisor:	Athanassios Boulis

Email of Supervisor:	athanassios.boulis@nicta.com.au

Name of Joint/Co-Supervisor:

Email of Joint/Co-Supervisor:

School:	School of Computer Science and Engineering

For CSE and EET Projects:	NICTA Project

Faculty Research Area (Theme):	Embedded Systems and Communications

School Research Area:	Networks, Sensor Networks, etc

Applicable to other Engineering schools/disciplines:

Abstract:	Castalia is an open source wireless sensor network simulator developed by NICTA with an active worldwide academic community. You will enhance some of Castalia's existing models based on measurements taken at NICTA. This work will give you great insight into large open-source software packages and the basics of wireless communication.

Research Environment:	You will be closely guided and supported by your academic supervisor having daily interactions. You will also have the opportunity to work along with other junior researchers in the research group. You will receive full support in terms or resources (computers, tools, IT services), enjoy the relaxed workplace, and take advantage of the seminars and workshops organized at NICTA to expand your academic and industry knowledge.

Novelty and Contribution:	Using real measurements to tune simulation models

Expected Outcomes:	Tested code for enhanced Castalia models.

Reference Material Links:	http://castalia.npc.nicta.com.au

Project Title:	Testing communication protocols in a real sensor network deployment

Name of Supervisor:	Athanassios Boulis

Email of Supervisor:	athanassios.boulis@nicta.com.au

Name of Joint/Co-Supervisor:

Email of Joint/Co-Supervisor:

School:	School of Computer Science and Engineering

For CSE and EET Projects:	NICTA Project

Faculty Research Area (Theme):	Embedded Systems and Communications

School Research Area:	Networks, Sensor Networks, etc

Applicable to other Engineering schools/disciplines:

Abstract:	The Roads and Transport Authority of NSW (RTA) together with NICTA are creating a sensor network test bed to monitor the structural health of bridges. The test bed will be functional by October 2009 and interesting research questions will need to get answered. One is the type of MAC used in the system. You will look at a few possibilities guided by your supervisor and you will have to implement them in the test bed using designs from other implementations or simulations. A detailed evaluation of the different possibilities will follow. Upon completion of this work a publication to a prestigious conference is very likely since these kinds of in-field studies are rare.

Research Environment:	You will be closely guided and supported by your academic supervisor having daily interactions. You will also have the opportunity to work along with other junior researchers in the research group. You will receive full support in terms or resources (computers, tools, IT services), enjoy the relaxed workplace, and take advantage of the seminars and workshops organised at NICTA to expand your academic and industry knowledge.

Novelty and Contribution:	Working with a real system deployed in a Sydney bridge

Expected Outcomes:	Implementation of communication protocols (MACs) and validation of their performance in a real system

Reference Material Links:	Contact supervisor

Programming Languages and Compilers

Project Title:	A Parallel Physics Engine

Name of Supervisor:	Manuel Chakravarty

Email of Supervisor:	chak@cse.unsw.edu.au

Name of Joint/Co-Supervisor:	Roman Leshchinskiy

Email of Joint/Co-Supervisor:	rl@cse.unsw.edu.au

School:	School of Computer Science and Engineering

For CSE and EET Projects:	School Project

Faculty Research Area (Theme):	Programming Languages and Software Engineering

School Research Area:	Programming Languages and Compilers

Applicable to other Engineering schools/disciplines:

Abstract:	For some years now, physical limitations prevent the individual processing units, or cores, of microprocessors to increase in performance. Instead, we see increasing numbers of cores per processor. To utilise multiple cores, programmers need to write explicitly parallel applications, and programming models to support this activity are an active area of research. Both data-parallel programming, where a particular operation is simultaneously executed on multiple data items, and functional programming, where the execution order is not statically fixed and side effects are rare, are widely recognised as promising approaches to exposing the increasing amounts of parallelism required by the growing numbers of cores per processor. The present project combines data-parallelism and functional programming in a case study from the performance-hungry application areas of interactive games, simulations, and virtual reality. The aim is to develop a purely functional, data-parallel core component of physical simulations, often called a physics engine, that computes the behaviour of physical objects, including their location, velocity, acceleration in response to forces, tracking collisions, etc. As part of the project, the TofR student will have to develop suitable data-parallel algorithms and implement these algorithms in the broad-spectrum functional language Haskell with data-parallel extensions.

Research Environment:	The project is hosted by the Programming Languages and Systems group of the School of Computer Science and Engineering and is part of a larger research effort to develop practical programming models for parallel computer hardware, specifically multicore CPUs and GPUs (Graphics Processing Units). The TofR student will work with one of the group heads and a postdoc and will gain first-hand experience in modern research in programming languages and parallel programming in an internationally widely recognised research group.

Novelty and Contribution:	Current physics engines are single-threaded or use a small number of concurrent threads. The aim of this project is to develop a massively data-parallel engine. This is both novel and necessary given the current development direction of computer architectures that increase resources for explicit parallelism, with very little -if any- increase in single-threaded performance. The concrete contributions of the project are algorithmic as well as with respect to parallel programming methodology.

Expected Outcomes:	(1) Implementation of a basic physics engine using data-parallelism in Haskell. (2) Performance benchmarks and a quantitative analysis of the exploited parallelism.

Reference Material Links:	This TofR project is loosely based on a past Google Summer of Code project and will benefit from the groundwork laid during that earlier project. Slides on data-parallelism in Haskell: http://www.cse.unsw.edu.au/~chak/papers/dp-in-haskell.pdf The Data Parallel Haskell project: http://www.cse.unsw.edu.au/~chak/project/dph/

Project Title:	Demand-Driven Dynamic Symbolic Execution for Detecting Bugs and Security Vulnerabilities

Name of Supervisor:	Jingling Xue

Email of Supervisor:	jingling@cse.unsw.edu.au

Name of Joint/Co-Supervisor:	John Potter

Email of Joint/Co-Supervisor:	potter@cse.unsw.edu.au

School:	School of Computer Science and Engineering

For CSE and EET Projects:	School Project

Faculty Research Area (Theme):	Programming Languages and Software Engineering

School Research Area:	Programming Languages and Compilers

Applicable to other Engineering schools/disciplines:

Abstract:	Testing is the primary way to check the correctness of software. Billions of dollars are spent on testing in the software industry as about half of the cost of software development is borne by testing. Software bugs are so prevalent and detrimental that they cost the USA economy alone an estimated $60 billion yearly. In addition, the historical record shows that these bugs are often fatal as they cause system crashes and frequently represent security vulnerabilities. Systematic dynamic test generation, by leveraging recent advances in symbolic execution and dyanmic test generation, is becoming increasingly popular because it can find bugs by automatically generating test cases without false positives. In the past few years, leading universities and labs such as Stanford and Microsoft have been conducting research in the area and produced a number of tools such as DART, EXE, CUTE, KLEE and SAGE. However, these tool suffer from a number of problems, including high analysis time and high false positives. In this project, we will continue our investigation that started initially as a summer project and later as a honours project by developing an demand-driven dynamic test generation tool in LLVM for detecting bugs and security vulnerabilities in C/C++ programs. By focusing our analysis on parts of the code that are the likely causes for bugs, our tool is expected to be more scalable with improved precision in practice.

Research Environment:	The student who will soon submit his honours thesis has developed a tool in LLVM that presently works on small programs. You will be expected to build your research on this and extend it so that the tool is capable of analysing million lines of code later. Our research group consists of seven PhD students working on a range of research problems in programming languages and compilers.

Novelty and Contribution:	We appear to be the first to reduce analysis time and improve analysis precision by applying slicing to symbolic execution. We also appear to be the first to focus on demand-driven symbolic execution.

Expected Outcomes:	An open-source tool is expected to be released once it is stable. A conference paper on the project is also possible. LLVM is an open-source production-quality compiler frameowork. The student is expected to develop significant software development skills with significantly improved understanding and knowledge about programming languages.

Reference Material Links:	1. Ryan Ericson's honours thesis. 2. LLVM: http://llvm.org/ 3. Symbolc execution and dynamic test generation: http://research.microsoft.com/en-us/um/people/pg/public_psfiles/ndss2008.pdf http://research.microsoft.com/en-us/um/people/pg/public_psfiles/pldi2005.pdf http://www.cs.berkeley.edu/~dmolnar/metafuzz-tr-draft.pdf

Project Title:	Dynamic Datarace Analysis for Concurrent Java Programs

Name of Supervisor:	Jingling Xue

Email of Supervisor:	jingling@cse.unsw.edu.au

Name of Joint/Co-Supervisor:	John Potter

Email of Joint/Co-Supervisor:	potter@cse.unsw.edu.au

School:	School of Computer Science and Engineering

For CSE and EET Projects:	School Project

Faculty Research Area (Theme):	Programming Languages and Software Engineering

School Research Area:	Programming Languages and Compilers

Applicable to other Engineering schools/disciplines:

Abstract:	As we enter the multicore era, concurrency is the next major revolution, after OO, in how we write software. However, multi-threaded programs are very difficult to write in terms of locks. Concurrent systems hang, fall prey to data races and deadlocks. A recent work, referred to as Goldilocks, demontsrates the feasibility of detecting dataraces dynamically in Java programs by ensuring that the analysis is both sound and precise. That work was implemented in a JVM but the source code wasn't made available in the public domain. Goldilocks represents an interesting research direction in dynamic datarace detection. In this project, we intend to implement it in IBM's Jikes JVM and improve its analysis efficiency by exploiting the object confinement information dynamically available in code.

Research Environment:	Our research group consists of seven PhD students working on a number of research topics in programming languages and compilers. There is also a postdoc with expert knowledge in OO, programming languages and formal semantics.

Novelty and Contribution:	The co-supervisor did some seminal work on object ownership and confinement about a decade ago. Both of us have been collaborating in the past few years on program analysis and concurrency for OO languages. Combining object confinement with dynamic analysis appears to be novel and is promising to reduce analysis overhead further.

Expected Outcomes:	A tool performing dynamic datarace detection for Java programs. A conference paper on this is possible if the object confinement information can be exploited to improve analysis overhead.

Reference Material Links:	1. Goldlocks: http://portal.acm.org/citation.cfm?id=1250762

Project Title:	The Semantics and Implementation of Cartesian Programs

Name of Supervisor:	John Plaice

Email of Supervisor:	plaice@cse.unsw.edu.au

Name of Joint/Co-Supervisor:	Blanca Mancilla

Email of Joint/Co-Supervisor:	mancilla@cse.unsw.edu.au

School:	School of Computer Science and Engineering

For CSE and EET Projects:	School Project

Faculty Research Area (Theme):	Programming Languages and Software Engineering

School Research Area:	Programming Languages and Compilers

Applicable to other Engineering schools/disciplines:

Abstract:	In Cartesian programming, programming takes place inside a multidimensional context, used as an index, where the number of dimensions is arbitrarily large. In the TransLucid programming language that I have designed and developed, equations are written in a form similar to those for differential equations, in which only the dimensions of relevance are described. In the same way that René Descartes's coordinate geometry allowed for the algebraisation of geometry, Cartesian programming makes it possible to have a single formalism in which to describe the entire development of a software system, with multiple heterogeneous components, in a fully declarative manner. The research will involve developing the Cartesian framework by studying known areas of computer science from a Cartesian perspective and developing the necessary software tools (interpreter, compiler, interfaces).

Research Environment:	The research will take place at the School of Computer Science and Engineering and will use school machines, along with software developed at the school.

Novelty and Contribution:	The research will develop innovative views of many existing aspects of computer science, along with new algorithms for parallel implementations of TransLucid.

Expected Outcomes:	Contributions to a production-quality interpreter for TransLucid.

Reference Material Links:	Toby Rahilly, John Plaice: A Multithreaded Implementation for TransLucid. IEEE COMPSAC 2008: 1272-1277.

Software Engineering

Project Title:	Adaptive exception healing in Cloud-based Web services

Name of Supervisor:	Jenny Liu

Email of Supervisor:	jenny.liu@nicta.com.au

Name of Joint/Co-Supervisor:	Anna Liu

Email of Joint/Co-Supervisor:	annaliu@cse.unsw.edu.au

School:	School of Computer Science and Engineering

For CSE and EET Projects:	NICTA Project

Faculty Research Area (Theme):	eResearch (Knowledge and Services Engineering)

School Research Area:	Software Engineering

Applicable to other Engineering schools/disciplines:

Abstract:	Cloud Computing (CC) is an emerging paradigm promising the on-demand provision of highly scalable resources and massive computing power to be consumed as services, such as Google App and Amazon EC. Running a medium to large sized software application can and never will be error free. This becomes even truer for components and services in highly distributed applications, where services rely on each other. Failure of vital system parts might turn out to be a threat to the correctness and reliability of the entire software system. The component oriented world addressed the common requirements for exception handling, logging and notification by reusable components or frameworks like Log4J, .Net Enterprise Library etc. However this approach is insufficient in a service oriented environment where exceptions occur from consumed services which are beyond direct control. In most cases the context of exception and the root cause remain untackled and the software fails. Consequently we see the need for a dedicated piece of software architecture, which is available at run-time and it is required to: • Categorize the exception • Provide detailed information revealing the root cause of failure • Indicate an appropriate healing strategy respective solution This project aims to design and prototype such an solution.

Research Environment:	TOR scholars will participate in NICTA's use-inspired research activities. The main and associated supervisors are quite experienced researchers with TOR projects. Since 2004, we have supervised two TOR projects every year. In 2008, one project won the best project award at NICTA ATP Laboratory. Students also have the opportunities in publishing outstanding research results out of the TOR project. We have several successful stories that TOR scholars published their work at top international conferences/journals and followed it up in their thesis projects. This project is hosted by NICTA Managing Complexity Group at Australian Technology Park. A dedicated computing lab with a cluster of high capacity computers are available for this project. Desk, workstation and computing facilities will be provided.

Novelty and Contribution:	This approach will enable easier and simpler tracking and responding to failure in CC applications. Thus enhancing applications running on CC resources with error management capabilities can support maintenance and developing personnel in order to provide faster and more reliable services to the end users.

Expected Outcomes:	• Integrating existing logging and monitoring tools with Cloud platform • A repository of exception category • Default implementation of healing strategies (could involve simple machine learning approaches) • A demo case

Reference Material Links:	http://aws.amazon.com/cloudwatch/

Project Title:	Business-Driven Selection of Web Services

Name of Supervisor:	Vladimir Tosic

Email of Supervisor:	vtosic@cse.unsw.edu.au

Name of Joint/Co-Supervisor:	Jacky Wai Keung

Email of Joint/Co-Supervisor:	Jacky.Keung@nicta.com.au

School:	School of Computer Science and Engineering

For CSE and EET Projects:	NICTA Project

Faculty Research Area (Theme):	eResearch (Knowledge and Services Engineering)

School Research Area:	Software Engineering

Applicable to other Engineering schools/disciplines:

Abstract:	Web services are distributed computing application components that use XML-based (Extensible Markup Language) technologies to implement the service-oriented architecture (SOA). They enable collaboration of heterogeneous computing systems distributed over the Internet, across corporate boundaries. They are already embedded in various products and services of all major computing companies. The Universal Description, Discovery and Integration (UDDI) is an industrial standard for describing businesses and their Web services and allows discovery and selection of Web services during run-time. However, UDDI describes only “what” Web services do and not “how well” they do it. As the number of Web services that offer similar functionality increases in the global market, their quality of service (QoS, such as response time, availability) and various metrics of business value (e.g., price, price/performance ratio, customer satisfaction) will become major competitive advantages. While several UDDI extensions specifying QoS information appeared recently, specification of business value is still relatively superficial. Students working on this project will extend an existing open-source UDDI implementation in Java with capabilities for publication and search of various business value metrics (and other relevant information). They will also develop novel algorithms for selection among similar Web services to maximise various business value metrics.

Research Environment:	The students will work closely with researchers at NICTA (http://www.nicta.com.au) in a friendly mixed-gender and multicultural team environment comprised of senior researchers and postgraduate students. For example, two female Ph.D. students with strong academic background and international industrial experience will be able to provide expertise and additional technical guidance, when needed.

Novelty and Contribution:	Members of our team have previously outlined some ideas how to extend UDDI with specification of business value (and other information, such as QoS). However, these ideas require improvements. Furthermore, without appropriate tools, theoretical innovations will not be very useful for software practitioners. The development of such tools is not trivial – before implementation, research questions should be answered. The developed tool will have significant advantages over the competition, because it will provide features (related to specification of business values and business strategies) that no current UDDI extension has. Some of the potential application areas are finance and e-government.

Expected Outcomes:	- UDDI data format extensions that enable specification of business value and other management information (with a previously developed format). - Extension of UDDI operations enabling publication and retrieval of various business value metrics (and other relevant information). - Extension of UDDI operations enabling search of Web services with particular characteristics (including characteristics in the standard UDDI and in the developed extension). - Novel algorithms for selection among similar Web services (found through a search query) to maximise various business value metrics. - Experiments (experimental data and their analysis) evaluating feasibility, usefulness, and performance characteristics of the implemented extensions.

Reference Material Links:	- http://www.w3schools.com/webservices/ws_intro.asp - http://www.cse.unsw.edu.au/~cs9322 - http://www.tutorialspoint.com/uddi/uddi_overview.htm - http://www.service-architecture.com/articles/index.html - http://ws.apache.org/juddi/ - http://www.nicta.com.au/people/tosicv/publications - Tosic, V. "On Modeling and Maximizing Business Value for Autonomic Service-Oriented Systems", keynote with invited paper in Proc. of Business Process Management (BPM) 2008 Workshops, Milan, Italy, September 1, 2008, Springer, Lecture Notes in Business Information Processing (LNBIP), No. 17, pp. 407-418. Preprint at: http://nicta.com.au/__data/assets/pdf_file/0007/17962/Tosic-invited-QSWS-08-FinalForLNBIP.pdf - For further information, email: Dr. Vladimir Tosic (‘vtosic’ at the CSE e-mail system) with Subject line “UNSW Summer Scholars”.

Project Title:	C/C++ Code Visualisation

Name of Supervisor:	Ralf Huuck

Email of Supervisor:	ralf.huuck@nicta.com.au

Name of Joint/Co-Supervisor:	Ansgar Fehnker

Email of Joint/Co-Supervisor:	ansgar.fehnker@nicta.com.au

School:	School of Computer Science and Engineering

For CSE and EET Projects:	NICTA Project

Faculty Research Area (Theme):	Programming Languages and Software Engineering

School Research Area:	Software Engineering

Applicable to other Engineering schools/disciplines:

Abstract:	Understanding software is a difficult task, this is in particular true for understanding source code. NICTA's Goanna project created an industrial strength bug checker that assists software developers in detecting defects and vulnerabilities automatically at compile time. However, for better understanding of the detected bugs a graphical display of source code is desirable. The goal of the summer project is to connect error traces in source code with graph visualisation software. This will enable a better understanding of why a bugs occurred and which parts of the code are impacted.

Research Environment:	The project is in collaboration with a team of international researchers, engineers, and students.

Novelty and Contribution:	The novelty about this summer project is the merger of visualisation tools with static program analysis results.

Expected Outcomes:	By the end of the summer project it is expected to have a prototype tool that takes current textual information such as control flow graphs, bug locations and error traces and maps those automatically to graph visualisation tools.

Reference Material Links:	http://www.nicta.com.au/goanna

Project Title:	Integrating mashp technologies with Web-based social network applications

Name of Supervisor:	Jenny Liu

Email of Supervisor:	jenny.liu@nicta.com.au

Name of Joint/Co-Supervisor:	Liming Zhu

Email of Joint/Co-Supervisor:	liming.zhu@nicta.com.au

School:	School of Computer Science and Engineering

For CSE and EET Projects:	NICTA Project

Faculty Research Area (Theme):	eResearch (Knowledge and Services Engineering)

School Research Area:	Software Engineering

Applicable to other Engineering schools/disciplines:	Biomedical Engineering Surveying & Spatial Information Systems

Abstract:	Mashup provides a way of forming new applications from existing Web content using APIs provided by different Web sites. Such a nature makes mashup a promising technology to deliver Web based social network application with rich information of various themes, such as participants' distribution, the categories of their preferences/information and etc. However, most mashup app only cover a single theme with a simple source of data format. Social network applications can contain large amount of information with complex connections with each individual. It is a challenging issue to present multiple themes of social applications. This project aims to deliver an integration solution to smartly retrieve and mashup information from different types of social network applications.

Research Environment:	TOR scholars will participate in NICTA's use-inspired research activities. The main and associated supervisors are quite experienced researchers with TOR projects. Since 2004, we have supervised two TOR projects every year. In 2008, one project won the best project award at NICTA ATP Laboratory. Students also have the opportunities in publishing outstanding research results out of the TOR project. We have several successful stories that TOR scholars published their work at top international conferences/journals and followed it up in their thesis projects. This project is hosted by NICTA Managing Complexity Group at Australian Technology Park. A dedicated computing lab with a cluster of high capacity computers are available for this project. Desk, workstation and computing facilities will be provided.

Novelty and Contribution:	Mashup and social network applications are the hottest topics on Web and service engineering. This project identifies a very challenging yet practical problem and aims to provide a novel solution for real world usage.

Expected Outcomes:	A number of information retrieval and mashup techniques will be developed using social network Web API and Google mashup libraries. A demo app will be produced.

Reference Material Links:	Contact supervisor

Project Title:	Software Engineering on Cloud Application Platforms

Name of Supervisor:	Anna Liu

Email of Supervisor:	annaliu@cse,unsw.edu.au

Name of Joint/Co-Supervisor:	Jenny Liu

Email of Joint/Co-Supervisor:	jenny.liu@nicta.com.au

School:	School of Computer Science and Engineering

For CSE and EET Projects:	School Project

Faculty Research Area (Theme):	eResearch (Knowledge and Services Engineering)

School Research Area:	Software Engineering

Applicable to other Engineering schools/disciplines:	Surveying & Spatial Information Systems

Abstract:	This project aims to examine the capabilities of the leading cloud application platforms: Microsoft Azure, Google App Engine and Amazon Web Services. We first build some sample applications on these platforms, and then evaluate the capabilities of these platform technologies in terms of performance and software engineering support. There is a good opportunity to build an innovative web 2.0 application using some of the latest technologies.

Research Environment:	The student has the option to work either at UNSW or NICTA (or both) and be supervised by senior researchers. There is also the opportunity to interact with professional engineers from Google, Microsoft and or web startups, as part of a research team at UNSW that collaborates heavily with industry partners.

Novelty and Contribution:	Cloud application platform is an emerging class of technologies that enables innovation on the web. The student will acquire valuable engineering skills and experience with these innovative platform technologies, and will contribute towards world class software engineering research that advances the state of practice for professional software engineers in Australia as well.

Expected Outcomes:	test applications, evaluation reports, presentation and report summarising findings.

Reference Material Links:	Google Microsoft Azure, Google App Engine, Amazon Web Services.

Project Title:	Web-based Process Mashup

Name of Supervisor:	Liming Zhu

Email of Supervisor:	limingz@cse.unsw.edu.au

Name of Joint/Co-Supervisor:

Email of Joint/Co-Supervisor:

School:	School of Computer Science and Engineering

For CSE and EET Projects:	NICTA Project

Faculty Research Area (Theme):	eResearch (Knowledge and Services Engineering)

School Research Area:	Software Engineering

Applicable to other Engineering schools/disciplines:

Abstract:	REpresentational State Transfer(REST) is the set of design principles behind the World Wide Web (WWW) which is evolving into powerful Web 2.0 mashup technologies.However, REST was mainly designed for transporting hypermedia data contents rather than facilitating process transactions. This has limited existing mashups to data mashup rather than process mashup. Students in this project will be exposed to the latest mashup technologies and improve them to support process-intensive systems.

Research Environment:	Students will work closely with researchers at National ICT Australia (ATP) in a very friendly team environment.

Novelty and Contribution:	Existing mashup approaches focus on data exposure and composition. The novelty of the project is on exposing process descriptions and enabling composition. Process fragments will be transported along with data to enable process-intensive systems.

Expected Outcomes:	Students will invent new micro-formats for process description exposure over the web. Students will implement a process exposure and coordination software prototype using existing web frameworks.

Reference Material Links:	Background reading: http://ieeexplore.ieee.org/xpl/freeabs_all.jsp?tp=&arnumber=4724571&isnumber=4724513 http://nicta.com.au/research/projects/armature

Theoretical Computer Science

Project Title:	Exploring SMT solvers for Software Analysis

Name of Supervisor:	Ralf Huuck

Email of Supervisor:	ralf.huuck@nicta.com.au

Name of Joint/Co-Supervisor:	Ansgar Fehnker

Email of Joint/Co-Supervisor:	ansgar.fehnker@nicta.com.au

School:	School of Computer Science and Engineering

For CSE and EET Projects:	NICTA Project

Faculty Research Area (Theme):	Programming Languages and Software Engineering

School Research Area:	Theoretical Computer Science

Applicable to other Engineering schools/disciplines:

Abstract:	Satisfiability Modulo Theories (SMT) problem is a decision problem for logical formulas with respect to combination of background theories such as reals, integers and various data structures. There are a number tools that solve these problems automatically. The goal of the summer project is to explore to which extend these SMT solvers can be used to analyze software such as C/C++ automatically. This project is targeted at students who have a good understanding of C/C++ code with a vivid interest in logics and/or mathematics.

Research Environment:	The summer student will be embedded in the Goanna project team that created an industrial strength bug checker assisting software developers in detecting defects and vulnerabilities automatically at compile time. The team consists of international researchers, engineers, and students.

Novelty and Contribution:	So far there is little work on integrating SMT solvers in existing real-life software analysis tools. The project will significantly improve the current understanding in the field.

Expected Outcomes:	By the end of the project a better understanding of SMT solvers, ideas for applying them to software source code, and estimations of their scalability are expected. These outcomes will contribute to the state-of-art research in this area.

Reference Material Links:	http://www.nicta.com.au/goanna http://en.wikipedia.org/wiki/Satisfiability_Modulo_Theories http://research.microsoft.com/en-us/um/redmond/projects/z3/

Project Title:	Soft Open Constraints in a Semi-Ring Framework

Name of Supervisor:	Michael Maher

Email of Supervisor:	Michael.Maher@nicta.com.au

Name of Joint/Co-Supervisor:

Email of Joint/Co-Supervisor:

School:	School of Computer Science and Engineering

For CSE and EET Projects:	NICTA Project

Faculty Research Area (Theme):	Intelligent & Autonomous Systems

School Research Area:	Theoretical Computer Science

Applicable to other Engineering schools/disciplines:	Mechanical & Manufacturing Engineering

Abstract:	The semi-ring constraint framework provides a way to define soft constraints - constraints that may be violated, but whose violation has a cost that must be factored in to the resulting solution. It has a solid theoretical foundation but considers only constraints of fixed arity. In practice, hard constraints of variable arity are used (for example, all_different(X1...Xn) constrains the values of X1,...,Xn to be distinct) and recently constraints that allow variables to be added during execution have been proposed. (These are called "global" and "open" constraints, respectively.) The aim of this project is to formulate global and open soft constraints within the semi-ring framework, preserving both the theoretical foundations of the semi-ring framework and the advantages of global and open constraints. The project will require the ability to perform abstract mathematical reasoning. Some familiarity with CSPs (Constraint Satisfaction Problems), such as in COMP3411 (Artificial Intelligence) or constraint programming as in COMP4418 (Knowledge Representation and Reasoning) is an advantage.

Research Environment:	You will be working closely with a senior researcher.

Novelty and Contribution:	Although the semi-ring framework is widely studied, no-one has investigated global and open constraints in this framework.

Expected Outcomes:	The expected outcome is a report and possibly a research paper describing the model of open global semi-ring constraints, including the statement and proof of its properties and algorithms for achieving local consistency within the model.

Reference Material Links:	The semi-ring framework is presented in the following papers: www.math.unipd.it/~frossi/jacm.pdf http://www.math.unipd.it/~frossi/sclp-toplas.ps.gz The formulation of soft open constraints in a different framework is given in http://www.cse.unsw.edu.au/~mmaher/pubs/cp/soggy.pdf

Web Services, E-Commerce, and other Web Technologies

Project Title:	An Empirical Evaluation of RDF Stores

Name of Supervisor:	Sherif Sakr

Email of Supervisor:	ssakr@cse.unsw.edu.au

Name of Joint/Co-Supervisor:	Boualem Benatallah

Email of Joint/Co-Supervisor:	boualem@cse.unsw.edu.au

School:	School of Computer Science and Engineering

For CSE and EET Projects:	School Project

Faculty Research Area (Theme):	Programming Languages and Software Engineering

School Research Area:	Web Services, E-Commerce, and other Web Technologies

Applicable to other Engineering schools/disciplines:	Biomedical Engineering Surveying & Spatial Information Systems

Abstract:	The vision of the Semantic Web has brought about new challenges at the intersection of web research and data management. One fundamental research issue at this intersection is the storage of the Resource Description Framework (RDF) data. The RDF data model has been designed as a flexible representation of schema-relaxable or even schema-free information. In RDF, all data items are represented in the form of (subject, predicate, object) triples, also known as (subject, property, value) triples. Several research eorts have proposed different techniques for storing and querying RDF datasets. These techniques can be broadly classied into two main classes: 1) Native RDF data stores 2) Relational stores for RDF data. The target of the project is to achieve an experimental comparison, analysis and benchmarking of the state-of-the-art of the stores. This experimental analysis could lead to identifying the strengths and weaknesses of each proposed approach and identify a set of metrics which could be used as a right indicator for determining the suitable storage technique for each dataset and its query workload.

Research Environment:	The student will work in an international group of PhD students, researchers and senior researchers in the Service Oriented Computing Research Group. Some literature review will be required to learn the basics of RDF and SPARQL query processing. Experimental analysis skills will also be acquired during the project activities. Project results will have a very good chance to be published in a good venue.

Novelty and Contribution:	- Benchmarking the-state-of-the-art of RDF stores. - An analysis of the strengths and weakness of the dierent techniques of RDF stores. - Identifying a candidate set of metrics for recommending the suitable storage schema of the RDF dataset and its query workload.

Expected Outcomes:	- This project will involve experimental analysis and benchmarking of the state-of-the-art of RDS query processors. - Literature scan of RDF storing and querying bibliography.

Reference Material Links:	- RDF Processing Bibliography: http://www.cse.unsw.edu.au/ssakr/RDFBiblio.htm - An introduction to RDF and SPARQL http://www.dajobe.org/talks/200603-sparql-stanford/ http://research.talis.com/2005/rdf-intro/ http://www.rdfabout.com/quickintro.xpd

Project Title:	Business Process Mining

Name of Supervisor:	Ghazi Al-Naymat

Email of Supervisor:	ghazi@cse.unsw.edu.au

Name of Joint/Co-Supervisor:	Boualem Benatallah

Email of Joint/Co-Supervisor:	boualem@cse.unsw.edu.au

School:	School of Computer Science and Engineering

For CSE and EET Projects:	School Project

Faculty Research Area (Theme):	eResearch (Knowledge and Services Engineering)

School Research Area:	Web Services, E-Commerce, and other Web Technologies

Applicable to other Engineering schools/disciplines:

Abstract:	The execution of business processes produces a massive size of event log data. This data can be exploited to discover efficient models of the business processes. The discovery process is called process mining, which allows for the analysis of business processes based on event logs. Many researchers have developed heuristics that focus on discovering process models based on an event log data. However, this project aims to apply one of the traditional data mining techniques to efficiently discover variation of well-organized process models. The technique should also provide a tool, which allows for the visualization of the discovered models. Examples of the traditional data mining techniques include: Association rules mining, clustering, and classification.

Research Environment:	The student will work in a team of researchers. A literature review is required to understand the fundamentals of process mining area. Implementing the proposed technique and conducting comprehensive experiments are required.

Novelty and Contribution:	1- Writing a comprehensive survey about process mining techniques. 2- Applying one of the traditional mining techniques to efficiently discover process models from event logs. 3- Presenting the results using an appropriate visualization tool.

Expected Outcomes:	A technique to discover variation of valid business processes from large event logs. A technical report that will be publishable as a conference or journal paper.

Reference Material Links:	W.M.P. van der Aalst, H.A. Reijers, A.J.M.M. Weijters, B.F. van Dongen, A.K. Alves de Medeiros, M. Song, and H.M.W. Verbeek. Business Process Mining: An Industrial Application. Information Systems, 32(5):713-732, 2007. http://prom.win.tue.nl/research/wiki/publications/start Han, J. & Kamber, M. Data Mining: Concepts and Techniques. The Morgan Kaufmann, 2001

Project Title:	Business-Driven Management of Web Services

Name of Supervisor:	Vladimir Tosic

Email of Supervisor:	vtosic@cse.unsw.edu.au

Name of Joint/Co-Supervisor:	Jacky Wai Keung

Email of Joint/Co-Supervisor:	Jacky.Keung@nicta.com.au

School:	School of Computer Science and Engineering

For CSE and EET Projects:	NICTA Project

Faculty Research Area (Theme):	eResearch (Knowledge and Services Engineering)

School Research Area:	Web Services, E-Commerce, and other Web Technologies

Applicable to other Engineering schools/disciplines:	Electrical Engineering & Telecommunications

Abstract:	Web services are distributed computing application components that use XML-based (Extensible Markup Language) technologies to implement the service-oriented architecture (SOA). They enable collaboration of heterogeneous computing systems distributed over the Internet, across corporate boundaries. They are already embedded in various products and services of all major computing companies. Management of information technology (IT) systems (including service-oriented systems) is the process of their monitoring and control to ensure correct operation, discover and fix problems (e.g., faults, performance problems), and achieve maximal benefits from the IT systems. Ideally, IT systems should be self-managing, or at least be managed with minimal human intervention. This is one of the goals of autonomic computing (effort lead by IBM). The current management solutions for service-oriented systems and business processes focus on optimizing technical quality of service (QoS, e.g., response time and availability), but provide inadequate treatment for optimizing business metrics, such as profit and customer retention. Business-driven IT management is an active area of research (lead by HP and IBM). Students working on this project will design, implement in Java, and test extensions of existing autonomic Web service management software with additional original support (new algorithms and data structures) for maximization of business metrics.

Research Environment:	The students will work closely with researchers at NICTA (http://www.nicta.com.au) in a friendly mixed-gender and multicultural team environment comprised of senior researchers and postgraduate students. For example, two female Ph.D. students with strong academic background and international industrial experience will be able to provide expertise and additional technical guidance, when needed.

Novelty and Contribution:	We have previously developed several Web service management tools with support for autonomic computing and/or business-driven management. This project will extend some of the existing software into a new autonomic business-driven management system for Web services. Before implementation can be done, some research questions should be answered. The developed system will have significant advantages over its competition, because in some aspects (maximisation of business values and enactment of business strategies) it will provide features that no current IT system management system has. Some of the potential application areas are finance and e-government. The developed solutions might lead to a patent.

Expected Outcomes:	- Software architecture that extends an existing Web service management system with additional modules relevant for maximisation of business value metrics relevant for company’s business strategy. - Detailed design of the extension modules. - Original algorithms that enable Web service management software to make management decisions that maximise business value metrics in complex situations, with minimal human involvement. - Implementation of these extensions in Java. - Experiments (collection of experimental data and their analysis) evaluating feasibility, usefulness, and performance characteristics of the implemented extensions.

Reference Material Links:	- http://www.w3schools.com/webservices/ws_intro.asp - http://www.cse.unsw.edu.au/~cs9322 - http://www.service-architecture.com/articles/index.html - http://www.research.ibm.com/autonomic/overview/ - http://www.nicta.com.au/people/tosicv/publications - Tosic, V. "On Modeling and Maximizing Business Value for Autonomic Service-Oriented Systems", keynote with invited paper in Proc. of Business Process Management (BPM) 2008 Workshops, Milan, Italy, September 1, 2008, Springer, Lecture Notes in Business Information Processing (LNBIP), No. 17, pp. 407-418. Preprint at: http://nicta.com.au/__data/assets/pdf_file/0007/17962/Tosic-invited-QSWS-08-FinalForLNBIP.pdf - For further information, email Dr. Vladimir Tosic (‘vtosic’ at the CSE e-mail system) with Subject line “UNSW Summer Scholars”.

Project Title:	Context-aware task sharing and management system for enterprise mobile workers

Name of Supervisor:	Helen Paik

Email of Supervisor:	hpaik@cse.unsw.edu.au

Name of Joint/Co-Supervisor:

Email of Joint/Co-Supervisor:

School:	School of Computer Science and Engineering

For CSE and EET Projects:	School Project

Faculty Research Area (Theme):	eResearch (Knowledge and Services Engineering)

School Research Area:	Web Services, E-Commerce, and other Web Technologies

Applicable to other Engineering schools/disciplines:

Abstract:	Enterprise workers are becoming increasingly mobile. In a large organisation, the need for effective coordination between such mobile workers arises in two fronts: (i) access to data/services and sharing of them, (ii) organising and scheduling tasks. One approach to effectively integrate the two fronts is to deliver data/service to the user in the context of the task to be performed (e.g., attending a meeting – ‘where’ is the meeting, ‘when’ is the meeting, ‘what’ is the meeting agenda, ‘who’ is attending the meeting, etc.). Tasks are often related to each other, forming what we refer to as ‘processes’. At any given time, the workers will be participating in multiple processes, each of them competing for the worker’s attention. Innovations in Location-Based Services (LBS) have demonstrated how location context could be utilised to deliver smarter applications to various domains (e.g., emergency response control, traffic management and advertising). We will apply context-awareness to deliver personalised service and task management systems in which the priority, urgency and importance of tasks are automatically managed based on a multitude of contexts such as location, time, people (collaborators) and dependencies between tasks. The system also determines available data and services to the user in the context of the task to be performed. Users then will be notified of such services (and data) that can assist their planning, organising, decision-making, or carrying out the tasks incisively.

Research Environment:	The successful candidate will work with an enthusiastic research team consisting of academic staff members from the school of computer science and engineering, postgraduate students and research assistants. There will be weekly progress meetings.

Novelty and Contribution:	Context-awareness in mobile environment still presents many challenges. This project particular focuses on understanding/modelling/reasoning about the context that is relevant in personal and collaborative environment. This understanding will lead to better provisioning of 'just-in-time' data and services delivery.

Expected Outcomes:	A prototype implementation of a task-oriented service delivery platform. A final written report.

Reference Material Links:

Project Title:	Context-sensitive Service Configuration and Delivery

Name of Supervisor:	Helen Paik

Email of Supervisor:	hpaik@cse.unsw.edu.au

Name of Joint/Co-Supervisor:

Email of Joint/Co-Supervisor:

School:	School of Computer Science and Engineering

For CSE and EET Projects:	School Project

Faculty Research Area (Theme):	eResearch (Knowledge and Services Engineering)

School Research Area:	Web Services, E-Commerce, and other Web Technologies

Applicable to other Engineering schools/disciplines:

Abstract:	A company offers a range of services that may create added value when they are delivered in an interrelated manner. Typical examples of these so-called “bundled services” are easily found: A conference venue booking might lead to discounted offers for catering services. A finance service company might recommend a home content insurance to someone just signing up for a home loan. However, in the world of competitive services market and discerning service consumers, a service provider needs a means to dynamically create and configure such a bundled service so that the various context sensitive information can be taken into account (e.g., if the customer already has contents insurance with the company, a different service might be offered instead). Also, creating a service bundle may mean selecting multiple services that cross organisational boundaries. Despite the apparent benefits, we are yet to see a working end-to-end solution to the problem. We envisage a service bundle provision environment in which both elementary and composite services are dynamically discovered, matched/combined and delivered to the users “at the right time” depending on the various business contexts present in a business process a user is engaged in.

Research Environment:	The successful candidate will work with an enthusiastic research team consisting of academic staff members from the school of computer science and engineering, postgraduate students and research assistants. There will be weekly progress meetings.

Novelty and Contribution:	The complex and ad-hoc nature of the bundled services call for intricate and sophisticated lifecycle management capabilities. This project aims to provide an end-to-end solution for authoring, delivering and managing context-aware service bundles targeting mobile users

Expected Outcomes:	Technology prototypes for modelling, implementing and managing context-sensitive service bundle as well as a formal written report.

Reference Material Links:	Value Webs: Using Ontologies to Bundle Real World Services, IEEE Intelligent Systems, 2004 Vol 19. Issue 4 pp 57-66

Project Title:	Distributed Decision Making Engine for Mobile Devices (iPhone)

Name of Supervisor:	Dr. Jacky Keung

Email of Supervisor:	Jacky.Keung@nicta.com.au

Name of Joint/Co-Supervisor:	Dr. Liming Zhu

Email of Joint/Co-Supervisor:	Liming.Zhu@nicta.com.au

School:	School of Computer Science and Engineering

For CSE and EET Projects:	NICTA Project

Faculty Research Area (Theme):	eResearch (Knowledge and Services Engineering)

School Research Area:	Web Services, E-Commerce, and other Web Technologies

Applicable to other Engineering schools/disciplines:

Abstract:	In large scale software systems, it is impractical to transfer all relevant information to a centralized system to make an informed decision at all times. Many decisions need to be delegated to local authorities. During this delegation, a decision-making template process with partial information could be transferred to the local authority. And then the local authority will combine it with additional local information to complete the decision making process. The DME engine will provide distributed decision-making capabilities using local information for portable and mobile devices. The focus of this work is to develop a software prototype on mobile portable devices such as an iPhone and other mobile endpoint devices. This allows information on the local device to be reused when needed, this is especially important for the future development of advanced and automated business process systems.

Research Environment:	You will be working with a team of experts in business processes, software architecture and machine learning at NICTA in a very friendly and dynamic environment. More importantly will gain academic and industrial exposure. Suitable for students interested in software design and implementation for web services and industry-scale software development.

Novelty and Contribution:	The contribution of this work is twofold: (1) A novel decentralised software architecture suitable for business process management systems. (2) User acceptance evaluation of the system will help in identify issues in its actual usage of the system on a daily basis. The experimental data derived from this work and the prototype provides a glimpse into the future of decentralised business process management system.

Expected Outcomes:	The student will gain in-depth knowledge in integrating components in software architecture for business process management systems. A simple software prototype of the system is to be developed by the candidate at the end the program.

Reference Material Links:	Apple iPhones/iPod Touch and Apple laptop/desktop will be provided for the development of an iPhone prototype system. Experience in Apple software development is essential, desirable to have experience in iPhone development. http://www.lixi.org.au http://www.nicta.com.au/research/projects/armature http://www.apple.com/iphone/

Project Title:	Distributed mining for complex patterns in large astronomical data

Name of Supervisor:	Srikumar Venugopal

Email of Supervisor:	srikumarv@cse.unsw.edu.au

Name of Joint/Co-Supervisor:	Ghazi Al-Naymat

Email of Joint/Co-Supervisor:	ghazi@cse.unsw.edu.au

School:	School of Computer Science and Engineering

For CSE and EET Projects:	School Project

Faculty Research Area (Theme):	eResearch (Knowledge and Services Engineering)

School Research Area:	Web Services, E-Commerce, and other Web Technologies

Applicable to other Engineering schools/disciplines:	Sciences – Maths, Physics, Chemistry

Abstract:	Researchers have been collecting large amounts of data on a continuous or periodic basis in many fields. Sloan Digital Sky Survey (SDSS) is vital example of such data; that is the most motivated astronomical survey project ever undertaken. SDSS Data Release 6 (DR6) contains more than 3.6 TB of data. Availability of such large amount of useful data is both an opportunity as well as a challenge for the application of data mining techniques to generate interesting information. The major reasons for the lack of such data mining applications in SDSS are the unavailability of data in a suitable format and its formidable size. This project will transform the data into co-location patterns using distributed techniques, to obtain additional galaxy/stars types from available attributes. These patterns will be used for finding undiscovered relations between galaxies/stars objects.

Research Environment:	The student will work with both the supervisors in order to develop the distributed techniques. We will look at distributing existing mining techniques. Necessary hardware will be provided to the student. The project may involve application of packages such as Hadoop so some experience with Java is recommended.

Novelty and Contribution:	Data mining techniques for large data collections are still evolving and so there is a large scope for new research. The project will produce a novel distributed algorithm for efficiently extracting a large amount of useful data from the SDSS catalog. This will help in finding interesting facts for the astronomy domain.

Expected Outcomes:	A software implementation of a distributed data mining algorithm for large data sets. A technical report that will be publishable as a conference or journal paper.

Reference Material Links:	Han, J. & Kamber, M., Data Mining: Concepts and Techniques, The Morgan Kaufmann, 2001 Ghazi Al-Naymat. Enumeration of Maximal Clique for Mining Spatial Co-location Patterns. Proceedings of the 6th ACS/IEEE International Conference on Computer Systems and Applications (AICCSA-08), Doha, Qatar. Mar 31st- Apr 4th, 2008. Pages (126-133) http://www.sdss.org/ Florian Verhein and Ghazi Al-Naymat. Fast Mining of Complex Spatial Co-location Patterns using GLIMIT. The 2007 IEEE International Conference on Data Mining (ICDM'07). Omaha NE, USA. October 28-31, 2007. Pages (679-684).

Project Title:	Flexible Modeling and Data Extraction of System Artifacts

Name of Supervisor:	Sherif Sakr

Email of Supervisor:	ssakr@cse.unsw.edu.au

Name of Joint/Co-Supervisor:	Boualem Benatallah

Email of Joint/Co-Supervisor:	boualem@cse.unsw.edu.au

School:	School of Computer Science and Engineering

For CSE and EET Projects:	School Project

Faculty Research Area (Theme):	eResearch (Knowledge and Services Engineering)

School Research Area:	Web Services, E-Commerce, and other Web Technologies

Applicable to other Engineering schools/disciplines:	Biomedical Engineering Surveying & Spatial Information Systems

Abstract:	Today every person has to manage a growing amount of artifacts. These data can be personal artifacts or business artifacts or combination of them. These artifacts can be: oce documents, emails, calendar data, pictures, database records, XML data,...,etc. Moreover, these data can be distributed over a huge range of storage devices like desktop computers, mobile phones, email servers, relational databases and WWW. The RDF data model has been designed as a flexible representation of schema-relaxable or even schema-free information. In RDF, all data items are represented in the form of (subject, predicate, object) triples, also known as (subject, property, value) triples. The target of this project is to design a flexible representation of system artifacts based on the RDF data model. The output of this project will be a tool which can generate the RDF data representation of system artifacts using a collection of wrappers. Given the data model of an artifact type X, the wrapper should extract the dened meta-data, data and relationships of artifacts of the type X and generates the RDF triples with respect to the artifact data model. The tool should be flexible enough to enable the end users to adjust these wrapper or even develop his own wrappers using a light-weight tools.

Research Environment:	The student will work in an international group of PhD students, researchers and senior researchers in the Service Oriented Computing Research Group. Some literature review will be required to learn the basics of RDF and data management of heterogenous artifacts. Data analysis skills will also be acquired during the project activities. Project results will have a very good chance to be published in a good venue.

Novelty and Contribution:	- Providing a flexible tool of dening and managing data models of heterogeneous artifacts. - A set of wrapper to extract the data from the system artifacts and generate RDF triples.

Expected Outcomes:	- This project will involve developing flexible wrappers to extract metadata and data from heterogenous artifacts. - Literature scan of modeling and managing heterogonous artifacts.

Reference Material Links:	Contact supervisor

Project Title:	Indigenous Community Search Engine

Name of Supervisor:	Cat Kutay

Email of Supervisor:	ckutay@cse.unsw.edu.au

Name of Joint/Co-Supervisor:	Helen Paik

Email of Joint/Co-Supervisor:	hpaik@cse.unsw.edu.au

School:	School of Computer Science and Engineering

For CSE and EET Projects:	School Project

Faculty Research Area (Theme):	Programming Languages and Software Engineering

School Research Area:	Web Services, E-Commerce, and other Web Technologies

Applicable to other Engineering schools/disciplines:

Abstract:	Knowledge available about Indigenous communities, their history and how government agencies view them, is important for these communities to plan their own development. Much of this information is available on the web or through government files, but must be interpreted in many ways to be accessible to most community organisations. The initial environment has been designed, and search environments such as Grokker are good examples of what is possible for the search interface. However the web applications to be developed will also need to provide the ability to extract and integrate different types of data. The research will look at the categories of searches that are used and develop an interface to enhance this searching for novice users. It will include the integration of novel applications for the integration of data from various formats in mashups.

Research Environment:	The project will involve working in a small team in a new research area. You will develop a prototype site for testing by selected users from community and government. You will be involved in working with users from the community to assess usage of the site and negotiating with government department for access to reports what are not previously online.

Novelty and Contribution:	At present there is very little resources online tailored to the needs of these communities, so issues such as general usability will need to be considered as well. The research is to be the basis of a proposed linkage grant with government agencies to provide this type of service to Indigenous communities.

Expected Outcomes:	At the end of this project, the candidate will provide a working prototype of a Web search and display system. The applications included will provide integration of information from sources such as: (a) pdf and word documents (b) GIS data (c) Contact details (d) News Blogs, and link these back to their original context. The research will lead to publication in Indigenous Knowledge and Community Development conferences in Australia, providing a social focus to IT research.

Reference Material Links:	http://www.grokker.com/ http://www.cities.org.au/DAA_Demo

Project Title:	Managment Framework for a Virtual Infrastructure

Name of Supervisor:	Srikumar Venugopal

Email of Supervisor:	srikumarv@cse.unsw.edu.au

Name of Joint/Co-Supervisor:

Email of Joint/Co-Supervisor:

School:	School of Computer Science and Engineering

For CSE and EET Projects:	School Project

Faculty Research Area (Theme):	eResearch (Knowledge and Services Engineering)

School Research Area:	Web Services, E-Commerce, and other Web Technologies

Applicable to other Engineering schools/disciplines:

Abstract:	Cloud computing is a novel paradigm that, in one form, allows users to lease compute and storage resources from providers for short intervals. Most of the compute clouds, such as Amazon EC2, are based on the use of virtual machines (VMs) to provide the requested resources. A VM is a complete representation of a physical machine which enables operating systems (and running applications) to be decoupled from the underlying hardware. At present, multiple such VMs can be run alongside each other on the same physical machine with minimal overhead, thereby allowing more efficient usage of physical resources. A common question for a cloud provider is how to maximise profit while ensuring that users obtain the capacity requested. The goal of this project is to develop resource management strategies for managing VMs so as to best serve the provider's purpose. Such strategies should take advantage of features in VM management frameworks such as Xen and VMware to suspend, migrate, expand or shrink VMs and will involve significant work at the system level.

Research Environment:	The student will work closely with a friendly group consisting of the supervisor and members of the SOC group in the school. The work will be carried out using the open-source Eucalyptus cloud computing framework developed by Univ. of California, Santa Barbara that provides the facilities for implementing strategies. Necessary hardware will be made available to the student.

Novelty and Contribution:	This is a fast changing research area with space for plenty of new contributions. One of these would be a novel resource management framework for VMs.

Expected Outcomes:	A modified version of Eucalyptus that can be used for further research into cloud services and a technical report that will be publishable in a good conference or journal.

Reference Material Links:	http://open.eucalyptus.com/

Project Title:	New Methods for Visualising Two-mode Networks

Name of Supervisor:	Michael Maher

Email of Supervisor:	Michael.Maher@nicta.com.au

Name of Joint/Co-Supervisor:	Lanbo Zheng

Email of Joint/Co-Supervisor:	Lanbo.Zheng@nicta.com.au

School:	School of Computer Science and Engineering

For CSE and EET Projects:	NICTA Project

Faculty Research Area (Theme):	Intelligent & Autonomous Systems

School Research Area:	Web Services, E-Commerce, and other Web Technologies

Applicable to other Engineering schools/disciplines:

Abstract:	Research Area: Constraint Programming (CP), Integer Programming (IP) and Graph Theory Crossing minimisation is one of the most fundamental problems in graph theory and has important applications in VLSI layout, graph visualisation and web-based rank aggregation. The aim of this project is to find good solutions for the two-clustered crossing minimisation problem which was recently used in visualisation of two-mode networks. Work will involve formulating the problem with CP and IP models; developing exact algorithms that integrate the strengths of CP and IP to approach optimal solutions; developing fast heuristic algorithms and evaluating them with real-world data sets.

Research Environment:	You will be working closely with senior researchers.

Novelty and Contribution:	This will be the first hybrid CP and IP algorithm for attacking the well-known crossing number problem. Research outcomes are extendable and applicable to a wide range of visualisation problems.

Expected Outcomes:	The student will gain basic knowledge in constraint programming and integer programming and advance to a higher level of graph theory. The student will also get familiar with some competitive optimisation tools and get to know the beauty of combinatorial optimisation!

Reference Material Links:	Graph Drawing: Algorithms for the Visualization of Graphs. G. D. Battista, P. Eades, R. Tamassia, I. G. Tollis. Prentice Hall 1999. Integrated Methods for Optimization, J. N. Hooker, Springer, 2007, ISBN 0387382720. Visualization of Genetic Networks: Edge Crossing Minimization of a Graph Drawing with Vertex Pairs (2000) A. Yamaguchi , H. Toh http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.21.5556

Project Title:	Service Oriented Architecture for e-Business Standards

Name of Supervisor:	Liming Zhu

Email of Supervisor:	limingz@cse.unsw.edu.au

Name of Joint/Co-Supervisor:	Ross Jeffery

Email of Joint/Co-Supervisor:	rossj@cse.unsw.edu.au

School:	School of Computer Science and Engineering

For CSE and EET Projects:	NICTA Project

Faculty Research Area (Theme):	eResearch (Knowledge and Services Engineering)

School Research Area:	Web Services, E-Commerce, and other Web Technologies

Applicable to other Engineering schools/disciplines:

Abstract:	Industry consortium have been developing e-Business standards using XML and business process modeling. Such standards inevitably have to be mapped to technology layers such as service-oriented architecture (SOA) and Web architecture. Deriving a flexible reference architecture and implementation from e-Business standards are not always straightforward. It involves multiple technical and not-technical factors and design trade-offs. This project will guide students to look into the state-of-art in e-Business standards, web service protocol stacks, service coordination mechanisms and Web 2.0 technologies. The work will contribute to the standardization body directly. Students will work with one of Australia's leading e-Business standardization body to solve real-world problems by inventing methods and implementing prototypes. Students may also have the opportunity to do additional work, which will be qualified for "Industry Training" required by the school. Most of students who are involved in this project previously have been offered a job in the financial industry.

Research Environment:	Students will work closely with researchers at National ICT Australia in a very friendly team environment. Suitable for students interested in software design, web services and industry-scale development.

Novelty and Contribution:	Existing e-Business standards focus on establishing shared data formats in the form of data models, XML schemas and common processes. There is gap between the data/process standards with implementation technologies such as SOA. This compromises the interoperability goal of such standard as different implementation interpretations lead to incompatible systems even using the same data format.

Expected Outcomes:	Students will map an existing data centric e-Business standard to a reference service architecture. Students will then generalise the mapping into an overall mapping methodology.

Reference Material Links:	http://www.lixi.org.au http://nicta.com.au/research/projects/armature

Computer Science & Engineering Research Areas

Related Projects

Computer Science & Engineering Projects

Algorithms