VIEWS: 19 PAGES: 11 POSTED ON: 3/13/2010
PF1 Guitar Hero is way too easy Supervisor Peter Farrell Students
PF1: Guitar Hero is way too easy Supervisor Peter Farrell Students Reynard Purnata Xuan Dao Chiang Satya Satya Mitchell Wong Project Description Guitar Hero, a game for the Playstation console, involves strumming a dummy guitar which contains 5 buttons whilst the game measures your skill. However, for skilled guitar players, this is far too easy. This project aims to create an interface such that players can use real guitars to play the game. This project will connect an electric guitar to a console using an ADC, a micro-processor and other minor hardware components. Taking into consideration synchronisation, timing and filtering issues, the game will be able to measure the game player’s actual guitar playing skills. MK1: Engineering devices educational tools for Signals and Systems Supervisor Margreta Kuijper Students Vijay Murali Dharan Jijoo Kim Project Description This project aims to show how signals and systems theory, the foundation of numerous engineering devices, can be used for educational purposes. To this end, a number of applications are given as examples. Image processing is investigated, with an image represented as a two dimensional array, and the Fourier Transform and various filters are applied to give different effects. Dual-tone and multiple-frequency, as well as audio processing are explored. Laplace transforms and their link to electrical circuit theory are investigated in designing an audio equalizer for enhancing music quality. MK2 - Codes for DNA computing Supervisor Margreta Kuijper Students Niedya Joeng Project Description Do you feel dissatisfied with today's computers' speed and performance? Are you frustrated with the fact that computers with great performance come in bulky sizes? Well fret no more! In the not so distant future, computers with high-end performances may come in test tubes. DNA computation requires problems to be formulated in codewords that take the form of DNA strands. Great care must be taken in choosing suitable DNA strands to ensure that a solution to the problem is generated. This project aims to give an overview of this exciting new area and investigate the suitability of ring codes over Z_4 in comparison to known additive cyclic codes over the field GF(4). MM2-1: Autonomous source localisation Supervisor Mark Morelande, Branko Ristic Students Alex Fan Christopher Leong Project Description With the number of nuclear devices in the world, and the threat posed by terrorist organisations acquiring them, the ability to identify and locate these hazards safely and efficiently is crucial. As conventional methods may place personnel in potentially dangerous situations, this project aims to tackle this problem through the development of autonomous vehicles which cooperatively work to localize such hazardous sources. For the purposes of development, a sound source will be used as a safe analogy to nuclear material. The main developments will involve simulating and eventually implementing source estimation, vehicle scheduling and communication algorithms on a robotic development platform that is purposely designed for research use. MM2-2: Autonomous source localisation Supervisor Mark Morelande, Branko Ristic Students Richard Wibisono Inno Tanaya David Gunawan Project Description With the emergence of hazardous danger particles in our environment, it is important that an Autonomous Source Localizer exists. In order to prevent humans approaching these dangerous particles, this platform can identify the particles and their positions. This project aims to enable the platform to find the source of the danger within some known area containing obstacles. To this end, an ultrasonic or light source is used to model a radioactive or dangerous source. An algorithm known as a particle filter is used to search for the particle. Further software and hardware development is employed in enabling the platform. PD1: Exploring Chaos in Semiconductor Lasers Supervisor Peter Dower, Peter Farrell and Kerry Hinton Students Koji Payne Daniel Thompson Project Description Lasers have many applications in modern society ranging from optical communication systems and surgery, to DVD burners. It is important to have a sound understanding of how lasers behave in order to maximise their performance, whilst using them safely and effectively. One laser element of particular interest is the semiconductor laser diode, which is a highly nonlinear device. The main aim of our project is to investigate whether these devices are capable of exhibiting chaos, and if so, formulating analog and digital control schemes to mitigate these chaotic and other undesired effects. PF-SP: PC sound card Supervisor Peter Farrel, Peter Dower Students Blair McCallum Brendan Westhorpe Project Description Current computer soundcards are designed to run inexpensive pc speakers, but when connected to a HiFi system, the sound quality can be poor. Our system aims to connect to a computer, bypassing the normal sound channels, convert from/to analogue and deliver high quality sound to an amplifier, or record with high quality from an analogue source such as a turntable. Our USB connected device will be external, eliminating the internal electrical noise of the pc. Data transformation will be between USB and I2S format, which can then be manipulated in a DSP to implement filters before finally being recorded or played. The aim is to have 24bit/ 96Khz sound recording and playback. TH2: Camera Tracking: Multi-Object Identification & Multi-Camera Tracking Supervisor Thomas Hanselmann, Ba-Ngu Vo Students John Li Peter Yang Project Description With the prevalence of surveillance cameras in modern society, the volume of video data available has exploded. As a result, online camera monitoring and offline video reviewing has become impractical. The aim of this project is to explore automated tracking and identification of multiple objects across multiple cameras. Our project will employ image processing techniques to extract specific features from within a video frame, and use them to identify and track objects across multiple frames both spatially and temporally. Autonomous methods of camera tracking will also be explored to avoid the need for manual calibration in deployment. TN1: Health monitoring for the elderly Supervisor Thas Nirmalathas Students Yana Helen Bittar Nadia Farhana Radzi Project Description Having home based health monitoring initiatives could help the aged and disabled lead a normal life in their familiar environment. With access to reliable monitoring technology, carers will be aided in effectively managing the condition of these people. Our aim is to build a gateway that can transmit data from health monitoring sensors such as glucose and oxygen level meters to a central processing unit via Zigbee. Here, data will be processed, analyzed and collated into a report. In the case of abnormal data, our system will alert carers and/or health care professionals via SMS or email. TN2: Automation of LED Lighting in Energy Efficient Homes Using Solar Energy Supervisor Thas Nirmalathas Students James Mackey Marcus Yong Project Description With the phasing out of candescent bulbs and the environmental problems associated with disposing of mercury in compact fluorescent bulbs, there is a need for a sustainable alternative. Recent advances in light emitting diode (LED) technology has led to LED’s offering a viable, efficient and environmentally friendly solution to home lighting. The aim of this project is to research, design and create an energy efficient, automated system of LED lighting in homes that is connected to solar power generation. The project will also analyse this system using data and models to investigate the feasibility of implementing this system in real world applications. TV1: Development of a Portable Gait Analysis System Supervisor Tharshan Vaithianathan, Subhash Challa Students Rahul Raghavan Aniket Sharma Project Description Recording human movement has allowed doctors to pick up and correct abnormalities just by assessing how a person walks. It has also allowed animators to create movies such as Happy Feet and Beowulf by mapping how a person moves and translating this into an animated figure. Conventional techniques for measuring an individual’s gait are expensive and cumbersome. This project aims to develop a simple to use, portable, yet inexpensive method of monitoring human movement. It involves integrating off-the- shelf MEMS sensors with wireless transceivers to create a small portable sensor node. This can be easily placed on the patient and their movement information sent wirelessly to a remote monitoring location. WS2: Coherent Optical Communication Supervisor Students Mansour Chamoun Dylan Cowen Project Description There is a new method of radio broadcasting being brought out, digital radio broadcasting. This method is superior to the old analogue method in almost every way, less interference and also allows playback and the sending of information such as the name of the song currently being played. Unfortunately receivers for this technology are very expensive, $150 for the cheapest. Our project aim is to design and build a low-cost digital radio receiver by throwing away most of the extra features and concentrating on only demodulating the audio information. JE1: Aluminium for Power Transmission Supervisor Jamie Evans Students Nick Edney Ramya Madhavan Project Description Current day power transmission relies heavily on copper as the primary material for distribution. Recent trends in the price of metals suggest that using a cheaper metal, such as aluminium, may provide a significant commercial advantage. This project aims to analyse the use of aluminium in preference to copper within the distribution chassis of a panel board. To this end, a theoretical analysis is carried out on the base properties of aluminium and copper, and the implications of these on the mechanical and electrical design is interpreted. An aluminium distribution chassis is tested and a commercial analysis is carried out to determine whether aluminium is indeed a viable and more competitive option. AK2: Low Cost ECG Monitor for Developing Countries Supervisor Ahsan Khandoker Jim Black Students Brian Walker Project Description An electrocardiogram (ECG) is a medical device used for monitoring a person’s heart beat. ECG’s are commonplace in our society, however many developing countries cannot afford enough of them. Our device will be low cost, and yet just as effective as expensive commercial equipment. It is hoped that hospitals in Mozambique will express an interest in the final product. A mobile phone will be used as a display, with the ECG connected via USB. Three probes attach to the patient - one on the wrist and one on either leg. Diagnosis will be made easier via user friendly visual interfaces on the phone which enable the user to easily freeze the display and backtrack. AK3: Sleep Apnea Screening Device for Home Monitoring Supervisor Ahsan Khandoker Students (Oscar) Yuanda Xu (Ken) Kian Hin Ang Project Description Sleep apnea refers to pauses in breathing during sleep. These pauses (apnea) occur repeatedly overnight, often resulting in sufferers experiencing sleepiness and fatigue during daytime. There is also an increased risk of heart conditions and other diseases associated with this sleeping disorder. Most sleep apnea cases currently remain undiagnosed because of the cost and limitations of clinical testing. This project aims to design a portable device for home diagnosis of sleep apnea, with monitoring and diagnosis in real-time. Using an algorithm developed last year for sleep apnea screening by means of ECG signals, hardware system implementation will be carried out using MATLAB Simulink and a DSP Board. BK1: Hands Free Mouse Supervisor Brian Krongold Students Daniel Cowen Project Description This project examines the potential for controlling a mouse pointer via a webcam by detecting eye movement. The possibility of controlling a pointer using head motion, by tracking dots on a person’s face, will provide primary insight into the viability of eye movement control. This project has exciting implications for the control of computer software by people with a disability as well as a viable future alternative to the hand controlled mouse. This project will be implemented using a C/C++ library entitled ‘openCV’. This project will involve interfacing with a webcam, processing the corresponding images, and then interfacing with Windows or an alternative GUI to control the pointer. BM1: Orbital Angular Momentum for Sensing and Communications Supervisor Prof. William Moran Prof. Peter Farrell Students Anthony Agosta Abang Ikhwan Abang Othman Ting Che Steven Chang Project Description In the early 1990s, it was recognized that electromagnetic radiation could possess orbital angular momentum (OAM) as well as spin angular momentum. This mechanical property of electromagnetic radiation can be manipulated and used for a wide range of applications. This new technology can lead to high-information density communications, sensor and signal processing applications, Optical Tagging and radar processing. Our project aims to produce beams of OAM photons at microwave frequencies and observe its divergence and reflective properties. We will first generate the desired OAM in the passive framework, by designing holographic gratings on a printed circuit board or similar structure. DG6: Speech enhancement using an auditory model Supervisor David Grayden Anthony Burkitt Students Aswin Wijetillake Gopi Pathmalinkam Project Description Most people would have experienced the frustration of loud background noise while talking on a mobile phone. Our project aims to develop an automatic speech enhancement algorithm which removes this noise from recorded or transmitted speech, hence improving the quality for the listener. This technology could then be used in various other applications such as the improvement of hearing aids. We will employ an auditory model, based on how the ear processes speech, along with further processing, such as formant tracking, to extract and enhance the most important properties of the speech signal while removing the noise. DG8: iHearU Supervisor David Grayden Students Zhongyu Qi Yanjun Ma Project Description Have you ever found it hard to talk to your friends in a noisy night club? What about struggling to hear a phone call at a football match? Our "iHearU" headset can help with these problems. Simply plug the headset into your mobile and turn your Bluetooth on. Much like helicopter pilots, you too can talk clearly in a noisy environment. Most importantly, it's small and portable. IHearU is a one-to-one walkie talkie system with noise cancelling technology, which exploits mobile phone Bluetooth capabilities, allowing people to talk for FREE. The beanforming technique is used to cancel noise at the microphone and adaptive noise-cancelling has been built into the earphones. EC2: Multiple Channel Cell Segmentation Supervisor Eric Dahai Cheng Subhash Challa Students Jiabao Su Zhenxin Zhang Project Description Microscopic cells tend to migrate over their lifetime. Traditional methods of detecting a cell and then tracking it through a series of images require a lot of human intervention, which is extremely time-consuming and expensive. Our project aims to design a Matlab program to detect and track cells in different types of microscopy images. To achieve this, we are going to perform a study into robust cell segmentation algorithms and perform simulations based on greyscale and CFSE images. The algorithms and simulation results will then be merged to create our program. The program must be robust to images of varying contrast ratios and noise levels. EW1: Word Recognition on a DSP Development System Supervisor Erik Weyer David Grayden Students David Quan Rebecca Wilson Project Description Our system aims to recognise certain words from a selected vocabulary, regardless of who the speaker is. There may be many applications for robust speech recognition, such as: control of electrical appliances around the home, word-display for the hearing impaired, or in multi-lingual translation. Using the "ADSP-BF548 EZ-KIT Lite®" DSP (Digital Signal Processor) Board (from Analog Devices), combined with signal processing techniques, speech analysis will be conducted in both time and frequency domains. This analysis will produce a system with robust algorithms which will recognise words from the chosen vocabulary, independent of speaker, and display the word or provide an indication that the word was successfully recognised. JE2: JIMIBOT Supervisor Jamie Evans Peter Farrel Joe Eden Students Tharindu Senarathna David Collins Kanupriya Sharda Project Description Like the rest of us, buskers have dreamed of taking a day off. To allow this luxury we aim to automate the guitar playing process. Coordinating the action of mechanical, electrical, and software components, ‘Jimibot’ will play an array of music to a skill level which would take a person many years to perfect. This project marks a new benchmark not only in robotic instrument performance, but also in music. A pneumatic cylinder matrix will emulate the finger positions along the frets and solenoids will carry out the plucking mechanism. This will be achieved using control theory, microcontrollers and the MIDI protocol. JE4: You’re not going anywhere! Supervisor Jamie Evans Students Garry Walpole Project Description This project is to develop a remote controllable braking system for a child’s bicycle. The system is to increase safety for small children on bicycles by enabling parents to remotely bring the bike to a stop without locking the wheels. The addition of the system should have minimal impact on the normal performance of the bike. The braking system will comprise a number of safety features including fail- safe brakes and anti-lock braking. Implementation will require innovations including; redesign of the brake calliper to an electromechanical unit, and the design of control system. The control system will respond to a variety of inputs, including radio remote control, and adjust the calliper accordingly. JM1: Developing Measurement Tools for Audio Amplifiers Supervisor Jonathan Manton Peter Farrel Students Shuyan Yu Hengjing Zhu Rui Su Li Lawrence Deng Project Description This project will investigate why two different audio amplifiers having similar, if not identical, 'specifications' can sound very different to the trained ear. Investigation into the existing methods of testing and evaluating different amplifiers will be conducted. Based on the technical data of two different amplifiers, a double-blind-test will be carried out. This will be used to identify if a normal audience can tell the difference in the quality and other amp specifications given different testing results. If possible, additional research will be done, exploring new ways of measuring or improving the current testing methods. JM2: Regularisation of Audio Distortion via Audio Feedback Supervisor Prof Jonathan Manton A/Prof Peter Farrell Students Tymur Tsarenko Project Description The project will aim to improve the performance of a typical home hi-fi system. With the use of microphones located around the room, the designed system will be able to compare the “heard” sound with the desired sound. The system will adjust the speaker output to improve the sound quality. To achieve this, a feedback control system will be designed and implemented to improve the output, particularly the bass. The actual audio response will be compared with the desired audio response and the control circuitry will endeavour to bring the actual audio response closer to the desired audio response. LJ1: Animal brain skull removal Supervisor Leigh Johnston; Nathan Faggian Students Anthony Vo Dayin Sun Project Description For neuroscientists, manually defining animal brain outlines from MRI scans can be a time consuming task. This project aims to develop a software tool to automatically extract the brain as a 3D image from the MRI scan. This project is in collaboration with the Neuroimaging group at the Howard Florey Institute. MATLAB will be used to develop the Brain Extraction Tool. The tool will discriminate the different structures of segmented mouse brain images such as bone, tissues and blood vessels. Image processing techniques based on Bayesian shape modelling will create a 3D model of the mouse brain, which will then be used in a deformable model algorithm to extract the brain. MC-SP: SoundBeam Supervisor Michael Cantoni Students Kyle Slater Project Description Conventional loudspeakers are great at filling a room with sound; however there are many applications that require the ability to control exactly where the sound can be heard in a room. This project aims to create a device capable of projecting audio as a beam of sound. The highly directional beam of sound will be achieved through the self- demodulation of modulated ultrasound, emitted from an array of ultrasonic transducers. This will involve modelling and simulating the non- linear propagation of high powered ultrasonic waves in air, and the design and construction of an ultrasonic array and driving/signal processing circuitry. PD-SP: Autonomous Video RC Car Supervisor Peter Dower Students Daniel Kabel Emad Chamie Benjamin Maxwell Project Description Have you ever played race simulators like Daytona at Timezone? Do you like racing RC cars? Our ‘Autonomous Video RC Car’ lets you race a real RC car using a PC racing wheel. A wireless camera gives you a driver’s seat view of the road ahead. We will also program the car to race by itself, and set a benchmark time for you to beat … if you can! A USB interface will be built between the PC and the cars’ transmitter. A control system will be designed to automate the car, using captured images from the wireless camera to determine the cars best response, and our interface to send the control signals. MC5: Modelling and control of a charge distribution network analogy for an irrigation channel Supervisor Michael Cantoni Students Filip Ivanovski Project Description Large networks of open-water channels are used throughout the world to distribute water to farms. The flow of water is set by gates situated along these channels. Regulating water-levels throughout the network provides a way of controlling the capacity of the channels and monitoring the supply of water to various distribution points. An electrical analogue to an open water channel will be designed, built and controlled. Charge will be distributed to a collection of capacitors, each representing the storage capacity of a section of channel, while also reflecting transport delays. The control objective will be to regulate the voltage across each capacitor. Decentralised and distributed feedback control strategies will need to be considered. SH1: High-Resolution Video using Compressive Sensing Supervisor Stephen Hanly Students Martin Ivanovski Sudhir Laxman Raskutti Project Description Digital videos are made up of many digital photos, or frames, displayed in quick succession. In many applications, such as security surveillance systems, physical constraints limit video performance in terms of the number of pixels and frame rate that the camera is able to capture. New techniques indicate that it is possible to overcome these hardware restrictions to extract high quality video from a limited amount of data. Our project aims to build a prototype video camera using coded aperture masks and novel compressive sensing methods to improve the performance of low resolution video systems. This approach should outperform existing super-resolution techniques such as frame-wise sampling and interpolation. XW2: Efficient background estimation and object detection in video surveillance Supervisor Xuezhi Wang Subhash Challa Students Dharen Nayer Matthew Klinko Project Description Say an emergency alarm rings and a building is evacuated. How do we know if everyone has left the building? Using a surveillance camera to count the number of people that have entered and left the building is one way. Our project seeks to automatically detect objects in video surveillance by removing the background. Separating dynamic objects, such as people, from a static background is an important preprocessing step in many computer vision applications. Accurate and efficient background removal is critical for interactive games, person detection and tracking, and even in people counting. This project challenges us to achieve fairly accurate results with minimal costs and resources.
Pages to are hidden for
"PF1 Guitar Hero is way too easy Supervisor Peter Farrell Students"Please download to view full document