Signal Processing Research Projects
Speech Processing for Cochlear Implants
(Supervisors: A/Prof. Eliathamby Ambikairajah and Dr Julien Epps)
Email: ambi@ee.unsw.edu.au; j.epps@unsw.edu.au
The speech processor used in cochlear implants analyzes the ambient speech signal, and converts it into a series of pulses that are used to stimulate parts of the cochlea. The objective of this summer project is to implement a cochlear speech processor in MATLAB, and developing an algorithm for the removal of noise using two microphones when the implant user is in a noisy environment. The enhanced speech will be re-synthesized in order to assess the quality that would be perceived by the implant user. At the completion of the project, the student(s) will have strengthened their signal processing knowledge, MATLAB skills and technical/research skills. This project can be taken either by an individual student or by a group of two students.
Automated Sound Signature Analysis
(Supervisors: A/Professor Eliathamby Ambikairajah, Dr Julien Epps and Professor Branko Celler)
Email: ambi@ee.unsw.edu.au ; j.epps@unsw.edu.au; b.celler@unsw.edu.au
The aim of this project is to research and implement a sound processing-based telecare system to enable elderly people to live independently in their own homes. Using a combination of time-frequency signal processing algorithms, the proposed monitoring system will identify clinically significant ‘signature’ parameters, and then automatically construct a dynamic model of the person’s activities. During the summer period, novel signal processing algorithms for joint multi-microphone based noise reduction will be developed. At the completion of the project, the student(s) will have strengthened their signal processing knowledge, MATLAB skills and technical/research skills. This project can be taken either by an individual student or by a group of two students.
Variable Speed Speech/Audio Playback
(Supervisors: Dr Julien Epps and A/Professor Eliathamby Ambikairajah)
Email: j.epps@unsw.edu.au; ambi@ee.unsw.edu.au ;
Time-scale modification of speech and audio signals is an interesting problem that goes well beyond simply resampling the signal, and has applications in packet loss concealment for voice over IP and browsing of speech/audio signals at a user-defined speed. Time scaling of speech and audio signals is also a feature of the recent MPEG-4 standardisation efforts. This project will look at simple techniques for time-scale modification of speech signals, based on models of the temporal masking effect of the human auditory system. This project will extend students’ MATLAB skills and signal processing knowledge, and will require some reading of technical papers. Suitable for one student.
Development of a simple Spatial Audio System
(Supervisors: A/Prof. Eliathamby Ambikairajah and Dr Julien Epps)
Email: ambi@ee.unsw.edu.au; j.epps@unsw.edu.au
Spatial audio is sound that has been processed (filtered) to give the listener a sense of the location of a sound source and the characteristics of a listening space. The human brain has developed complex auditory cues to decipher the approximate location of origin. The time delay between the arrival of a sound between the left and right ears, known as Inter-aural Time Difference (ITD), is the most significant auditory cue. When a sound signal is filtered by accurate Head Related Transfer Functions (HRTFs) and sent to the listener's two ears using headphones, the synthesized sound is experienced as a virtual source at the desired location in space. The goal of this project is to create a simple spatial audio system, involving recordings of sound by placing two small microphones in the left and right ears; comparing the recorded sound to the original sound and studying the HRTFs. The filtered signal using HRTFs will be delivered to headphones to create the perception that the sound comes from sources outside the head.
Stress Assessment by Combined Speech and Physiological Analysis
(Supervisors: Dr. David Shi (NICTA) and Mr. Ronnie Taib (NICTA))
Email: david.shi@nicta.com.au; ronnie.taib@nicta.com.au
Automated detection of stress while operating a human computer interface (HCI) can be done by obtrusive means e.g. physiological sensors, or implicit means e.g. changes in speech characteristics. Based on existing software (mainly in Java) and hardware pieces this project will involve the design and development of a software monitoring, analysis and logging interface for physiological and speech characteristics, all for the purpose of assessing stress levels in real-time. Working with a team of senior researchers and engineers, the student will gain skills in advanced technologies such as speech processing, user interface design, real-time Java programming. Using combined speech and physiological analysis to assess stress is at the leading edge of stress management and usability evaluation for HCI. The work carried out in this project fits into a broader cutting edge research project at NICTA focusing on cognitive load measurement through multimodal interaction. This project has direct applications in situations of high load, such as road or air traffic control. The NSW RTA is a current partner.
Developing an FPGA-based Smart Camera for HCI Applications
(Supervisors: Dr. David Shi (NICTA) and Mr. Ronnie Taib (NICTA))
Email: david.shi@nicta.com.au; ronnie.taib@nicta.com.au
An FPGA is a high performance computing platform widely used in networking, image/video processing and other high processing power demanding applications. This project aims at researching into efficient ways of integrating various image processing modules into an FPGA based intelligent camera platform. The project will involve HDL and C/C++ coding, debugging and real-time testing. Working alongside a senior researcher, the student will have a great opportunity to gain skills in advanced FPGA technologies and computer/machine vision, he/she will also learn how to do research on real-time video processing and build an intelligent video camera and gain insight into embedded imaging system. The reason a smart camera is ‘smart’ lies in the fact that there exists inside the camera a processing unit which performs application specific information processing. Using an FPGA to build a smart camera, from scratch, is a challenging piece of work. The primary function of the camera to be built – gesture recognition – has a wide range of applications. The outcome of the project will be a smart camera built from scratch – from image sensor to display.
Multi-channel Active Noise Cancellation
(Supervisors: Alex von Brasch and Deep Sen)
Email: a.vonbrasch@unsw.edu.au, dsen@ee.unsw.edu.au
Active Noise Cancellation (ANC) is the technique, common is aircraft cabins, were an out-of-phase audio signal is generated to cancel the noise of the enclosure. ANC principles are well established for the single channel case for low frequency noise (like engine noise). The interest here is on extending these principles to noise of higher frequencies and wide bandwidths, using multiple canceling sources. This project would involve experimentation in the anechoic chamber to develop a system to cancel chatter noise (background speech), by processing microphone signals in MATLAB to generate the noise canceling feeds. The project is suited for students with some background in signal processing, and a strong interest in audio. It is suitable for one student.
Data and Mobile Networks Research Projects
Turbo Coding on high noise channels
(Supervisors: Alex von Brasch)
Email: a.vonbrasch@unsw.edu.au
Turbo codes, along with LPDC codes, are the most successful forward error correcting codes for wireless communications, and are set to become standards of the mobile communication networks of the future. However, the performance of turbo codes on harsh channels is seen as been inferior to established techniques, particularly outer Reed-Solomon codes with an inner convolutional code. The aim of this project is to establish the conditions and a paradigm for when the performance of Turbo codes deteriorates, by extensive simulation and experimentation across a variety of channels. The project will involve simulations in MATLAB and experimentation with the Sundance software defined radio (SDR). This project is aimed at students with an interest in mobile communications, and would require a moderate degree of computer programming skills. It is suitable for a single student.
Space-time processing and coding for wireless communications
(Supervisor: A/Professor Jinhong Yuan)
Email: Jinhong@ee.unsw.edu.au
Next generation wireless and mobile communications require transmitting and receiving multimedia information with high quality and throughput. Current research shows that by employing multiple-transmit and multiple-receive antennas it is possible to achieve very high data rate transmission in rich scattering wireless environment. Signal processing and coding combined with multiple antennas are called space-time techniques. The project is proposed to design transmitted signals with the space-time techniques. The candidate will work with a senior researcher and postgraduate research students. The work could include space-time modulation, wireless channel simulation, receiver design, etc. Programming is required to implement the designed schemes. Further enquiry, please contact Dr. Jinhong Yuan at 9385 4244 or via email.
Capacity analysis for wireless relay communications
(Supervisor: A/Professor Jinhong Yuan)
Email: Jinhong@ee.unsw.edu.au
Wireless relay has been proposed for the enhanced data rates for future
cellular radio systems to guarantee reliable packet data communications. It employs multiple terminals to cooperatively transmitting information from source to destination. By exploiting the user diversity (the different
channel conditions of the users), this technique can improve the system
capacity while reducing the transmission power with potential applications
for many wireless systems, for example, military battle field, mining field,
etc. The project is proposed to design relay strategy attempting to maximize user throughput while satisfying a maximum power constraint and analysis its capacity. The candidate will work with a senior researcher and postgraduate research students. The work could include forward error correction, relay strategy design, etc. Programming is required to implement the designed schemes. Further enquiry, please contact Dr. Jinhong Yuan at 9385 4244 or via email.
Systems and Control and Biomedical Research Projects
Real Time Controller Implementation using NI LabView
(Supervisors: A/Prof. Tim Hesketh and Dr Ray Eaton)
Email: T.Hesketh@unsw.edu.au; r.eaton@unsw.edu.au
LabView provides facilities for design and control of instrumentation. Design and programming is undertaken on a Windows platform, which supports user interfaces etc. However, real-time execution depends on code being downloaded to real-time embedded processors of various capabilities. Of particular interest is the ability to support Digital Signal Processor development boards. While these are designed primarily for signal processing applications, control requirements are very similar in many instances. The mathematics supporting discrete and continuous time systems should be able to exploit the capabilities offered by a digital signal processor and associated instrumentation. Such a system could be capable of relatively high performance as an embedded controller. The skills required will be applied mathematics and systems theory, the development of ability to program using the block-diagram coding associated with LabView, and the ability to undertake careful experimentation and testing of systems using laboratory processes.
Energy Systems Research
Condition monitoring of power system equipment
(Supervisor: Dr. Toan Phung )
Email: toan.phung@unsw.edu.au
High-voltage power transformers are critically important in the operation of transmission and distribution networks. Such transformers are usually constructed using oil/paper insulation systems. Transformer failure is mostly caused by the gradual breakdown of the insulation material within it. Recovery Voltage Measurement (RVM) is a non-intrusive and non-destructive technique that can be used to evaluate the condition of the oil/paper.
An automated RVM system was recently developed by the HV laboratory of UNSW. It is a computer-based measurement and control system that utilises National Instruments data acquisition hardware and LabView software. A measurement cycle consists of four phases: charging, discharging, measurement and relaxation. The test object is charged for a specific duration at a fixed voltage (up to 1kV) and subsequently discharged. The voltage across the test object is then measured for a specific duration. The voltage reading recorded is known as the Recovery Voltage. From this recovery voltage graph, the peak voltage and the time at which it occurs are recorded. The peak voltage is then plotted against the charging time. This process is repeated for various charging times. A series of peak voltages against the various charging times would then reveal a unique polarization spectrum for that test object and its insulation condition.
This project will involve the use of this system to evaluate the insulation properties of biodegradable oil. This type of insulant is increasingly being used in distribution transformers as an alternative to mineral oil. Also, the system is to be further developed to enable measurement of the polarisation and depolarisation currents (DPC). Expertise in computer interfacing and programming is essential together with practical skills in electronic circuit design.
Microelectronics and Quantum Computing Research Projects
Modelling of charge-based qubit devices for solid state quantum computation
(Supervisor: Professor Andrew Dzurak)
School of Electrical Engineering & Telecommunications & Centre for Quantum Computer Technology
Email: Andrew Dzurak – a.dzurak@unsw.edu.au
The Centre for Quantum Computer Technology (see www.qcaustralia.org) is focused on the fundamental physics and technology of fabricating a revolutionary silicon based solid state quantum computer prototype. Quantum computers represent the next generation technology in computing and electronics. Through manipulation of quantum states, they offer parallel processing power and capacity in applications of commercial and national significance. A silicon based quantum computer will be realised by embedding phosphorus ions regularly in a lattice of pure silicon. Various strategies have been developed which use the nuclear spin, electron spin or electron charge associated with the phosphorus ions to store the quantum information.
This research project will suit a student with interests in the following areas: solid-state electronics, physics, computer modelling.
The student will work in a team environment at the Centre amongst PhD students, postdoctoral researchers and academic staff.
The project will focus on the development of computer models for prototype quantum computer devices, using the commercial software ISE TCAD on our Centre’s Linux cluster. It will involve liaison with other students and researchers to identify specific modelling problems and then a period of model development, followed by a written report on the project.
Up to two positions will be available.
Wide band transimpedance amplifiers for optical to electrical conversion for optical interconnects.
(Supervisor: Professor Chee Yee Kwok)
Email: cy.kwok@unsw.edu.au
Metal interconnects in VLSI chips for high speed operation is reaching its fundamental limits because of the inductive, capacitive and resistive effects at such high frequencies. Optical interconnects is actively investigated as a solution to this road-block in the semiconductor industry where processes are projected to operate in access of 20GHz by 2015. One of the requirements for optical interconnects to work is to be able to convert electrical signals to optical signals and optical signals to electrical signals. This project focuses on high speed optical to electrical conversion circuits in the form of a transimpedance amplifier. The project involves designing the circuit with state-of-the-art Cadence CAD software using silicon on sapphire(SOS) technology where substantial amount of circuit simulation using Cadence will be done. Project is suited for potential first class honours student with a liking towards analog integrated circuit design. The project is part a larger research project on MEMS based optical interconnects for Systems in a Package funded by the ARC.
Microelectronic circuit design for biomedical implants
(Supervisor: Dr Torsten Lehmann)
Email: tlehmann@unsw.edu.au
The aim of this project is to design a low-power CMOS integrated circuit for
use in biomedical implants such as a bionic eye. Researchers at UNSW has been working on the bionic eye for many years and the actual circuit to be designed in this project depends on the current state of the research as well as the interests of the students involved; examples include neural response amplifiers, high-efficiency power supplies, and accurate AD/DA converters. The project consists of three parts: 1) familiarisation with integrated circuit design techniques, 2) design specification, and 3) design of the circuit using professional CAD software. The project involves working with senior researchers in the field. At the end of the project, the participants will have strengthened their skills in microelectronic circuit design and the use of CAD tools, and will have developed research skills in preparation for their thesis work.
It is preferred that two students should jointly undertake this project, however it can be tailored for a single student if necessary.
Microelectronic circuit design for quantum computing
(Supervisor: Dr Torsten Lehmann)
Email: tlehmann@unsw.edu.au
The aim of this project is to design a low-power CMOS integrated circuit for use a very-low (sub-1K) temperature operation. Low-temperature integrated circuit design is an emerging field of research at UNSW and the actual circuit to be designed in this project depends on the current state of the research as well as the interests of the students involved; examples include pulse generators, low-noise amplifiers, fast AD/DA converters. The project consists of three parts: 1) familiarisation with integrated circuit design techniques, 2) design specification, and 3) design of the circuit using professional CAD software. The project involves working with senior researchers in the field. At the end of the project, the participants will have strengthened their skills in microelectronic circuit design and the use of CAD tools, and will have developed research skills in preparation for their thesis work.
It is preferred that two students should jointly undertake this project,
however it can be tailored for a single student if necessary.
Photonics Research Projects
Fluorescence in quantum-dots doped polymers
(Supervisor: Professor Francois Ladouceur)
Email: f.ladouceur@unsw.edu.au
The inclusion of Quantum Dots -- or nanocrystals -- into a polymer matrix can lead to the controlled emission of fluorescent light. These hybrid materials can then be used to create organic light emitting devices (OLED) with a wide range of application. This project would look at the synthesis of such materials and if time permit to their inclusion into functional devices. The project would be suitable for one student.
[Top of Page]
