|
|
|
Bias in LocalizationProject Code: SEaCS_03Supervisors:Prof. Brian Anderson, ANU and NICTA, and Dr Brad Yu, ANU and NICTAOutline:Can one determine an emitter/target location without using expensive global positioning systems (GPS)? Yes, fundamental theoretical advancement has enabled us to do that, with a small number of anchor sensors whose location information is available and without knowing the exact locations of the other sensors. The applications are of great interest to both civil and defence authorities. The technology is constrained by the noise and bias exist in the measurement and the method, which then lead to a degree of inaccuracy in the localization. We create a framework to analyse this general problem and would like to invite the applicant to work out solutions to a number of specific scenarios. Prerequisites:The student should be familiar with MATLAB programming. Contact:E: Dr Brad YuImplementation of a 3D Spatial Audio System using head Related Transfer Function (HRTF) modellingProject Code: AE_01Supervisors:Ms Wen Zhang (PhD Student) Associate Professor Thushara Abhayapala (ANU)Outline:3D spatial sound techniques are now commonly used in multimedia PCs and home theatre systems become. There are 3D sound systems using only 2 speakers to enhance the sound field effect, while maintaining the original sound characteristics. Two major components in the 3D sound are the stereo image enhancement and the head related transfer function (HRTF). The stereo image enhancement is targeting for widening stereo images by manipulating the stereo information, i.e., the difference between the left channel and the right channel. It can be implemented through some sophisticated gain controls, in audio frequency sub-bands. The HRTF’s are usually obtained from measurements on people (or on dummy head). It contains all the relevant spatial cues and describes how a given sound wave input is filtered by the diffraction and reflection properties of the individual body shape before the sound reaches the listener’s eardrum. Hence HRTF’s can be used for changing the directional cue of a sound source by applying an appropriate transfer function.Goals of project:The aim of this project is to develop a 3D digital audio system by combining the above two techniques through some proprietary tuning. The work is based on our newly-developed continuous HRTF model, which provides appropriate HRTF smoothing for 3D moving sound effects. Prerequisites:The student participating in this project is expected to have some knowledge in signal processing and MATLAB programming. The student will work closely in a research team with Wen and Thushara, and will have access to an acoustic lab to have fun. References:W: Audio and Sound Design in Virtual EnvironmentsW: Head-related transfer function Contact:E: Wen ZhangE: Thushara D. Abhayapala W:Thushara D. Abhayapala Source localization and separation with binaural Head-related transfer functionsProject Code: AE_02Supervisors:Ms Wen Zhang (PhD Student) Associate Professor Thushara Abhayapala (ANU)Outline:The acoustic environment poses at least two important challenges. First, people must localize sound sources using a variety of binaural and monaural cues; and second people can separate sources into distinct auditory streams (the “cocktail party problem”). While binaural cues include intra-aural time and level difference, the primary monaural cue is the spectral filtering introduced by the head and pinna via the head-related transfer function (HRTF), which imposes different linear filters upon sources arising at different spatial positions. Goals of project:It is generally understood the binaural HRTFs provides information about the location of a sound source and the HRTF plays an important role in source separation. In this project, we aim to address both challenges by exploiting the binaural HRTF to separate spatially localized acoustic sources in noisy environment. Prerequisites:The student participating in this project is expected to have essential prior knowledge in signal and systems and basic digital signal processing. The student should also be capable of programming in MATLAB. The student will work closely in a research team with Wen and Thushara and have access to an acoustic lab to have fun. References:W: Sound localizationContact:E: Wen ZhangE: Thushara D. Abhayapala W:Thushara D. Abhayapala 3D Microphone array for directional soundfield recordingProject Code: AE_03Supervisors:Ms Aastha Gupta, A/Prof Thushara Abhayapala Applied Signal Processing Group, InfoEng, RSISEOutline:Arrays of microphone structures are capable of processing acoustic signals to extract useful spatial information of the three dimensional surroundings. These are important in many three dimensional applications such as beamforming, direction of arrival estimation, spatial soundfield recording etc. Recently, we have proposed a new 3D microphone array structure that consists of circular microphone arrays. The goal of this project is to construct a 3rd order array using miniature microphones and implement the related signal processing techniques. Prerequisites:An enthusiasm to learn about new, exciting and inspiring research. An ability to tinker with electronics. Above all, a self motivation that this is the project for you and this program is a fulfilling way for you to spend your summer. Reference:T.D. Abhayapala, A. Gupta, "Non Spherical Microphone Array Structures For 3d Beamforming And Spherical Harmonic Analysis", submitted to the 11th International Workshop on Acoustic Echo and Noise Control download paper Contact:E: Thushara D. AbhayapalaE: Ms Aastha Gupta Are You About To Die?Project Code: AE_04Supervisors:A/Prof Thushara Abhayapala Sandun Kodituwakku Applied Signal Processing Group, InfoEng, RSISEOutline:The surface electrocardiogram (ECG) is used as the gold standard for diagnosing heart related disorders in medical practice. The ECG is a multi-component signal and consists of multiple waves, which can be traced back to different chambers (atria, ventricles) and different nodes (sinus-atria node, atria-ventricular node) of the heart. It is important to decompose the ECG into its original components in order to isolate and analyse some of the disorders. One of the major challenges is to extract the atrial activity from the surface ECG, as it has substantially low amplitude within the ECG. Extracting the atrial activity is highly useful in analysing atria related disorders like Atrial Fibrillation and Atrial Flutter. Several approaches have been taken for this purpose, including average beat subtraction, template matching and spatiotemporal cancellation of ventricular activity. Recently, principal component analysis and independent component analysis techniques have been used to estimate the atrial activity. All the above mentioned methods have pros and cons and none of the methods was capable of achieving a totally satisfactory solution. Goals of project:The goals of this project are to quantitatively analyse existing techniques for different ECG data banks and find a better method of extracting atrial activity by combining different techniques. For a creative student, there is a scope for coming up with novel ideas and developing his/her own algorithms as well. At the end of the project, the final algorithm will be implemented in Matlab, with the possibility of publishing the work. Prerequisites:The student is expected to have good mathematical background, programming skills and problem solving skills. The ability to program in Matlab will be helpful but not essential. The student is not expected have any prior knowledge in medicine or biology. Student’s gain:Bio-medical signal processing is an emerging research field and the student will be exposed to real life engineering problems in medicine and biology. The project has important clinical implications, and thus novel work will be well rewarded in both medical and engineering research communities. Contact:E: Thushara D. AbhayapalaE: Sandun Kodituwakku W: summerProj0809 Optimizing Channel Estimation for Higher Data Rate in Wireless Communication SystemsProject Code: WSP_01Supervisors:Dr. Parastoo Sadeghi, Dr. Tharaka Lamahewa, and Mr. Sean Zhou Department of Information Engineering, RSISE, ANUOutline:In wireless communications, the signal transmitted through the wireless environment (i.e. the channel) experiences distortion in its amplitude and phase. The receiver needs to have an accurate estimation of what has been done by the channel in order to decode the signal. Conventional wireless networks usually spend a considerable amount of transmit resource on channel estimation. For example, GSM system sacrifices about 22% of the data transmission time to perform channel estimation. However, this results in a direct reduction in the actual data rate. Therefore, optimizing the amount of resource spent on channel estimation is crucial for high data rate transmission, which has been an active area of research in recent years. In this project, the student will gain a good understanding on transmission design and performance analysis for future wireless communication systems. After an introduction to the topic and relevant background studies, the student will choose either of the following options:
Student’s gain:The project will provide an opportunity for the student to work in a research team with researchers in the Research School of Information Sciences and Engineering (RSISE) at ANU. It is expected that the outcome of the research be submitted as a technical paper to a national or international conference for publication. Prerequisites:Familiarity with wireless communications and Matlab will be a benefit. References"Pilot symbol transmission for time-varying fading channels: an information-theoretic optimization," Proc. Int. Conf. on Signal Processing and Commun. Syst., Dec. 2007. W: Download Paper Contact:E: Parastoo SadeghiE: Tharaka Lamahewa E: Sean Zhou
Modelling and Control of Autonomous FormationsProject Code: SEaCS_01Supervisors:Prof. Brian Anderson, ANU and NICTA, and Dr Brad Yu, ANU and NICTAOutline:How do swarms of agents, like fish or unmanned autonomous vehicles, manage to move in a formation, split a formation, merge formations, follow a leader, change formation and so on without colliding, and without a master device to drive them? Problems like this are the fascinating province of cooperative control, with various defence and civil applications. Goals of the project:This project aims to build and analyse different modelling tools for Autonomous Formations, which consist of several agents (you can think of simple robots). One typical mission we want to control this formation to merge with another formation, to join as a single rigid formation for greater functionality. Cooperativeness, communication efficiency and computational requirement are the key metric to evaluate the design. Note there is a chance for your design to be demonstrated with the testbed of up to 10 robots (see other robotic projects offered). Prerequisites:Visualization and testing are important in this project, and development of simple graphical tools may be needed. Interested students should be capable of programming in C/C+ or Matlab. Contact:E: Dr Brad YuRobotic SWARM on a testbedProject Code: SEaCS_02Supervisors:Dr Brad Yu, ANU and NICTA, and Mr. Eric Hou, ANUOutline:Inspired from biological swarms of bees, birds, ants, fish, etc., many researchers today study efficient coordination of teams of aerial, ground and underwater vehicles for various cooperative missions. A typical mission may involve forming a certain team structure (formation), maintaining or changing such a formation, intelligent guidance for cohesive motion of the whole team, etc. without a master device to drive the vehicles in the team. Efficient coordination of vehicles in such missions has various defence and civil applications. Goals of the project:This project aims to utilize a testbed to build SWARM robotic applications involving teams of up to 10 wheel robots such as E-puck or Surveyors SRV-1. The student will normally develop algorithms in Matlab (separately to this task) and validate using Webots Simulator before implementing with the testbed. Interested students should be capable of programming in Matlab and C/C++. Contact:E: Dr Brad YuFrom graphs to matricesProject Code: SEaCS_04Supervisors:Prof. Brian Anderson, ANU and NICTA, and Dr Brad Yu, ANU and NICTAOutline:Starting from real world observations about formations and abstracted them using graphs, be it undirected or directed, one obtains a direct view of the communication and information architecture among different nodes/agents/vertices. But how do we compare two graphs modelling two similar-looking formations. We want to tell ultimately which one is better than the other. Various matrices, but not those of topological sort like an incidence matrix, are used to best capture the tiny difference that are visually negligible but potentially differ drastically. Goals of the project:This project aims to study the transposition of the knowledge between two fundamental fields of mathematics. A number of former developed graph theoretical results are to be verified or advanced using theories and procedures in which matrices are extensively involved. Prerequisites:Interested students should be capable of programming in Matlab, and have good knowledge in linear algebra and discrete mathematics. Contact:E: Dr Brad YuHow to Make a Robot WalkProject Code:CV_01Supervisor:Dr Roy FeatherstoneOutline:Walking is something that humans do very easily, but robots find rather hard. When a human walks, every footstep is different: the exact timing and placement of each foot's landing, and the thrust delivered by each foot on the ground, are modified at each step in order to maintain balance, accommodate uneven terrain, adjust overall speed and heading, and so on. Robots can do this too, but not very well. This project is a simulation study of robot walking. It will take into account the dynamics of the robot's body (treated as a rigid-body system), the dynamics of the contacts between feet and ground, and the control system that makes the robot walk. It will also model the energy consumed during walking, so that the energy-efficiency can be measured. The objective will be to experiment with different control systems, different walking styles and different designs of robot legs, in order to see which ones work best. Prerequisites:To do this project, you will need a basic knowledge of rigid-body dynamics (e.g. Newton's laws of motion, and concepts like momentum, angular velocity and moment of inertia). You will also need a basic knowledge of how physical processes (like motion) are simulated on a computer: the process is modelled as a differential equation, and the computer solves the equation by numerical integration. (Essentially, it's an initial-value problem.) You should also be acquainted with the concept of feedback control systems, or, failing that, be acquainted with the use of feedback in electronic circuits. The project will make use of Matlab and Simulink, and it is desirable that you be familiar with both. This project offers you the chance to learn about advanced methods in dynamics from the world's leading expert on robot dynamics, to learn how a control system can make a robot walk (without falling over), and to learn how to model, simulate and analyse complex electro-mechanical systems, like robots. Personal tuition will be provided by Dr. Featherstone on these topics. Contact:E: Dr Roy FeatherstoneW: Dr Roy Featherstone Robots with Muscles made of Shape Memory Alloy WireProject Code:CV_02Supervisor:Dr Roy FeatherstoneOutline:Wires made from shape memory alloy (SMA) have the property that they can be easily stretched when cold, but will contract forcefully (like a real muscle) when heated. If we arrange the wires in antagonistic pairs, then they can operate the joints of a robot arm in much the same way that antagonistic pairs of real muscles operate the joints of a human arm. Of course, it's not quite as simple as that. If we want fast and accurate motion (which we do) then we need a good control system; and that is what this project is all about: designing control systems to obtain fast and accurate motions from antagonistic pairs of SMA wires. To see what we have already achieved, take a look at our Project Web Page. Your task would be to build on our success by working with our hardware and performing practical experiments to implement and test new control systems. This project would suit someone with a talent for working with electronic hardware and performing practical experiments. It offers an opportunity to learn how control theory works in practice, to learn about SMA and the world's best control systems for SMA actuators, and to work with high-performance real-time hardware and software (dSPACE hardware and Simulink). You might also have a chance to participate in the design and construction of a new robot to replace the antograph robot shown on the project web page. Prerequisites:Students undertaking this project would be expected to already have some experience of practical control systems. Prior knowlegde of SMA and prior experience of using dSPACE hardware are not required, but some knowledge of Matlab and Simulink is desirable. You would be working under the supervision of Dr. Featherstone, who will teach you all the required specialist knowledge. Contact:E: Dr Roy FeatherstoneW: Dr Roy Featherstone W: Project Web Page Boosting for Efficient Face DetectionProject Code:CV_03Supervisor:Dr Chunhua Shen, ViSTA, NICTA/RSISEOutline:Human face detection is of wide research interests. It is often necessary to detect them as a first stage of a processing chain for many tasks like face recognition, expression analysis, video surveillance, etc. Modern face detectors are usually based on Boosting techniques. Boosting is used to select relevant features from a huge number of candidates. At the same time, Boosting builds a strong classifier based on those selected features. Current detectors are very good at finding upright well-lit faces seen from the front, but significantly less good at tilted and profile faces under difficult lighting. Also, running a full classifier at all points of an image is computationally expensive. The project provides hands-on experience of building a sophisticated modern face detector using Boosting methods. Prerequisites:This is a challenging project. The ideal candidate would be a student who has extensive programming experience in C/C++. The project will begin by building a cascaded classifier with Haar-like features. References:
Contact:E: Dr Chunhua ShenGesture recognition for conducting the ARTEMIS orchestraProject Code:CV_04Supervisor:Dr Nick Barnes, DCS, InfoEng, NICTA VISTAOutline:The ARTEMIS initiative is a European Union project that is developing embedded systems of which NICTA is a member. The ARTEMIS orchestra is a competition to develop a robotic orchestra. Part of that includes recognising gestures from a conductor. Specifically tracking a conductor baton and adapting playing appropriately. This project would develop computer vision algorithms to track a conductor's baton and interpreting that information for a robotic instrument player. The algorithms would be designed to run on an embedded computing platform. Students should have an interest in computer vision and experience in C/C++. Information about the competition can be seen at:
Artemis Orchestra Including about NICTA's first prize winning entrant from this year. Contact:E: Dr Nick BarnesRemove camera shake from a single image (photo)Project Code:CV_05Supervisor:Dr Hongdong LiOutline:Unless you are using an expensive professional digital camera with VR (vibration reduction) functionality, camera shake during exposure often lead to objectionable image blur and ruins many photographs. Conventional image post-processing methods typically assume overly simplified forms for the camera motion path during camera shake. In this project we will explore and experiment on a newly-introduced method to remove the effects of camera shake (or motion blur) from a single seriously blurred image (ref: [1][2]). References:
Prerequisites:The student would be expected to have knowledge of calculus and linear algebra, Matlab or C/C++ programming experience. Having basic knowledge of signal processing or linear system or control will be beneficial. The student will work closely on a daily basis with the supervisor and postgraduate students. (RSISE) Contact:E: Dr Hongdong.LiW: Dr Hongdong.Li Photo re-focusing using depth-map estimated from a single imageProject Code:CV_06Supervisor:Dr Hongdong LiOutline:In professional portrait photograph it is often desired to have a sharp foreground with a blurred background. This is traditionally achieved by using a profession-grade DSLR camera with big-aperture lenses. However, many inexpensive point-and shoot digital cameras are often unable to implement such a selective blur effect, because of their relatively small lenses. In this project, we will study and implement a new method for digital photo post-processing to simulate such a selective blur effect from a single photo obtained by an inexpensive digital camera. At the hart of the method is a computer vision algorithm that estimates the raw depth-map estimation from a single image. References:
Prerequisites:The student would be expected to have knowledge of calculus and linear algebra, Matlab or C/C++ programming experience. Having basic knowledge of signal processing or linear system or control will be beneficial. The student will work closely on a daily basis with the supervisor and postgraduate students. (RSISE) Contact:E: Dr Hongdong.LiW: Dr Hongdong.Li Where am I ? ---Computer Vision-based localization (or, a camera-phone based GPS)Project Code:CV_07Supervisor:Dr Hongdong LiOutline:‘Location’ is a significant type of information that we are using for every day life. The extremely successful stories of GPS system and Google-Map are evidences of its significance. This project aims at developing a new computer vision-based technique for location recognition, using a common mobile camera-phone. This project borrows idea from a recent ICCV student contest at http://research.microsoft.com/iccv2005/Contest/ where a detailed project description can be found (ICCV 2005). However, for this summer project we will develop a simplified prototype software system based on ANU’s campus map. The main technical components that will be implemented include image feature detecting, image matching, and image database retrieval. Prerequisites:The student is expected to have some experiences with Matlab or C/C++ programming, have basic knowledge of linear algebra, and feel excited about doing image processing and computer vision research. The student will be working closely on a day-to-day basis with the supervisor and PhD students. (NICTA VIBE) Contact:E: Dr Hongdong.LiW: Dr Hongdong.Li Camera tracking with dynamic programmingProject Code:CV_08Supervisor:Dr Hongdong LiOutline:Feature matching and camera tracking are the process of extracting accurate tracks of 3D features observed in 2D video. Points of interest are indicated with a single mouse click in one frame of the video, and the desired output of the tracker is the location of the point’s 2D projection in every frame of the sequence. This technique is very useful in producing special video effect in a movie or generating realistic scenes in computer video games. In this project we will study a new algorithm based on dynamic programming ([1]) . This new algorithm works very fast and reliably. It is based on the k-d trees and dynamic programming —both are well-known techniques in computer science and software engineering. Reference:[1] Interactive Feature Tracking using K-D Trees and Dynamic ProgrammingBuchanan et.al., CVPR 2006. Prerequisites:The student would be expected to have knowledge of calculus and linear algebra, Matlab or C/C++ programming experience. Having basic knowledge of image processing will be beneficial. The student will work closely on a daily basis with the supervisor and postgraduate students. (NICTA VIBE) Contact:E: Dr Hongdong.LiW: Dr Hongdong.Li Smart Home CameraProject Code:CV_09Supervisor:Dr. Li Cheng, CRL NICTA, Adjunct Research Fellow, RSISE, ANUOutline:We deal with the practical problem of recording and storing live video capture over longtime (eg. days or weeks) as well as video summarization. In other word, we aim at an intelligent video footage storage and management system based on event detection (eg. motion detection) to eliminate unnecessary storage of situations where eg. the scene is essentially unchanged, as well as summarizing the visual events. This project is important in many applications: for example, to monitor the home remotely when the owner is away on holidays and he/she is still able to inspect his/her home from time to time from internet. Goals of project:
Prerequisites:Good programming skills in C/C++ Contact:E: Dr Li ChengHierarchical Image SegmentationProject Code:CV_10Supervisor:Dr. Li Cheng, CRL NICTA, Adjunct Research Fellow, RSISE, ANUOutline:In analogy to text that has a semantical hierarchy as character, word,sentence, etc, image can also be naturally represented using a hierarchical structure, where a node in this hierarchy indexes a group of nearby pixels, and to traverse upward to its parental node in the hierarchy corresponds to a higher level 'semantical' granularity. The task is to automatically build a hierarchical (or tree) structure for an input image (eg. an outdoor image) where each level of the hierarchy reveals a sensible segmentation of this image of certain granularity. Goals of project:
Prerequisites:Good programming skills in C/C++ Contact:E: Dr Li ChengThe Robot Walking ChallengeProject Code:CV_11Supervisor:Dr Roy FeatherstoneOutline:This project is a variant of the project "How to Make a Robot Walk". The challenge is this: you are given a dynamics simulator, a model of a walking robot and a model of slightly uneven terrain, and the task is to create a control system that will make the robot walk over the uneven terrain without falling over. To make the challenge hard, the robot is blind, and there is a 0.25 second time delay between the control system and the actuators. In other words, the actuators (muscles) will not respond to a command coming from the control system until 0.25 seconds after it has been sent. The purpose of the time delay is to mimic the kinds of delays that are present in biological systems (e.g. humans), and to get you thinking about how a control system (or a brain) could cope with these delays. Goals of project:This project is aimed at students who want to test their problem-solving skills on a hard problem that will require a great deal of imagination and ingenuity to solve. For other details of this project, see "How to Make a Robot Walk", project CV_01. Contact:E: Dr Roy FeatherstoneW: Dr Roy Featherstone Stereo Visual-sensing system for a Small-scale FlyerProject Code:R&CV_01Supervisor:Dr. Jon KimOutline:Low-cost visual sensing systems on small-scale flyers have much potential in future robotic applications such automated search and information gatherings. Goals of project:The aim of this project is to implement an aerial visual mapping system using two wireless web cameras with fish-eye lenses and a ground-based image acquisition system. The depth extraction or disparity map will be constructed by processing the image data in Matlab environment. The successful results of this project will lead to a more challenging problem of Simultaneous Localisation and Mapping (SLAM) fusing vision and inertial data on a small-scale flyer which has never been tackled. Prerequisites:This project requires experiences on Linux-based programming and Matlab image processing. Contact:E: Dr Jonghyuk KimData fusion for Range and Visual-homography InformationProject Code:R&CV_02Supervisor:Dr. Jon KimOutline:In recent times, laser range scanners based on Photonic Mixer Devices (PMD) camera technology have become commercially available. PMD represents a fast and accurate method to scan the 3D environment within the camera's field of view. The camera provides 3D range point clouds and monochrome image. There has been work to register the 3D range point clouds using the 3D ICP (Iterative Closet Point) algorithm. One problem is it does not work well if the environment is flat and monotonous such as road or wall. Goals of project:The aim of this project is to utilise the visual homography from the road image collected previous experiment in the car-park to enhance the existing ICP performance. Prerequisites:This project requires experiences on Matlab image processing. Contact:E: Dr Jonghyuk KimInvestigation of optical motion cues for stable teleoperation of robotic vehicles.Project Code:CV_12Supervisors:Nick Barnes (NICTA), Robert Mahony (CECS)Outline:Computer vision has the potential to provide a robust and rich exteroceptive sensor for control of autonomous vehicles. Optical flow is a measure of image motion as observed by a camera. Under certain lighting and scene assumptions, optic flow has the potential to provide stabilising feedback for the control of robotic vehicles. In this project, the student will trial a range of optic flow-based motion cues for teleoperation of semi-autonomous robotic vehicles. Amongst the motion cues trialled will be spherical divergence, differential optic flow, and total optic flow. The motion cues and resulting control algorithms will be implemented on the InsectBot holonomic robotic platform in NICTA. Contact:E: Dr Nick BarnesE: Dr Robert Mahony
Computational and experimental analysis of failure of composite structuresProject Code: DE_01Supervisor:A/ Prof Shankar KalyanasundaramOutline:A sustainable future for human society necessitates using advanced materials for various applications including aerospace, bioengineering and automotive industries. In order to use these materials effectively in these application, one needs to characterize their mechanical properties and their failure. One of the aims of this project is to use an advanced photogrammetric system (the only one of its kind in Australia) to characterize the material behavior. The system uses two CCD cameras and monitors the deformation of materials in real time as the forces acting on these materials are increased and ultimately the material fractures. A pattern is sprayed on the material and the system (ARAMIS) keeps track of the deformation of the entire structure through deformation behavior of the pattern. A high speed version of this system was used to elucidate the failure of the last space shuttle accident. As a part of this project, the student will develop a thorough working knowledge of this system and learn to use it effectively to characterize the failure behavior of the composite materials. This system is widely used around the world in various research laboratories and provides the student a good understanding of an extremely useful research tool in experimental mechanics. The computational analysis of this project will involve using finite element analysis with different failure theories built in it. Experimental results will then be used to validate the effectiveness of different failure theories. Contact:E: A/ Prof Shankar KalyanasundaramProperties of n-type ribbon-grown silicon for solar cells
Semi-Markov Conditional Random Fields for Sequence SegmentationProject Code: SML_1Supervisor:Prof. Wray Buntine, CSL, NICTA SMLOutline:Automatic paragraph segmentation (APS) refers to the problem of automatically segmenting written text into paragraphs. This is useful to make computers understand natural language. The recent model of semi-Markov Conditional Random Fields (semi-CRFs) allows incorporating features of a whole paragraph such as paragraph coherence, and transitions within a paragraph can be non-Markovian. Besides, semi-CRFs only incur a small amount of extra computational cost compared with conventional CRFs. Recent research has shown that semi-CRFs are excellent in solving named entity recognition and other related natural language problems. Therefore, it is both worthwhile and challenging to explore and extend the use of semi-CRFs in paragraph segmentation. Goals of the project:
Prerequisites:Good programming skills in C/C++, with gcc and Linux. Good knowledge in dynamic programming and mathematics. Reference:[1] Sunita Sarawagi & William W. Cohen. Semi-Markov Conditional Random Fields for Information Extraction. NIPS 2004. W: Postscript[2] Qinfeng Shi, Yasemin Altun, Alex Smola, and S. V. N. Vishwanathan. Semi-Markov Models for Sequence Segmentation. In Proceedings of the 2007 Conference on Empirical Methods in Natural Language Processing (EMNLP), 2007. W: ShiAltSmoVis07 Contact:E: Prof. Wray BuntineWeb Interfaces for a Semantic-based Search EngineProject Code: SML_2Supervisor:Prof. Wray Buntine, CSL, NICTA SMLOutline:We are building a semantic-based search engine for Parliament House to use on some of their content such as Hansard, or the Acts for Parliament, or the Journals. The content is placed in structured XML, and we are to develop novel user interfaces for semantic-based search. First, we need to augment the actual content with some kind of semantic markup, such as topics, names, or such. Then, we need to let the search users take advantage of this information. This is an open-ended project that will allow you to experiment with different web-based interfaces. The semantic markup will be arranged by our group, but you will have to make use of it in a novel interface. Goals of the project:
Prerequisites:Good programming skills in some web-oriented framework such as Java, Python, etc. and in Linux, and on web-based GUIs such as in AJAX. General knowledge of search, XML, web languages and scripting. Contact:E: Prof. Wray BuntineDefining PDDL++Project Code: S_02Supervisor:Dr. Patrik Haslum, DPO Group (ANU) & Managing Complexity (NICTA)Outline:In recent years, the Planning Domain Definition Language (PDDL) has become the de facto standard formalism for specifying AI planning problems. PDDL is highly expressive in principle (it can model the dynamics of any finite state transition system, and with recent extensions also some kinds of timed/hybrid transition systems), but viewed as modelling language it does leave a lot to be desired. Goals of the project:The aim of this project is to create a more convenient (easy to use) modelling language for planning problems, that can be efficiently compiled into PDDL, and to implement this compilation. Such a modelling language could borrow concepts or constructs from modern programming languages, for instance object orientation (inheritance and polymorphism) or templates, but also build on concepts from research in automated planning, for instance finite domain variable representations or generic types. The envisioned language is primarily intended for modelling finite and deterministic planning problems (so called "classical planning"), and the compilation to generate problems in the corresponding subset of PDDL, but could also be extended to capture problems with metric time and resources, "soft goals", or uncertainty (which would then be compiled into one of the extended PDDL fragments). The compilation could also be extended with more advanced forms of problem analysis and simplification. Prerequisites:This project requires good software design and programming skills (it is, after all, about creating something close to a programming language), and some familiarity with elementary logic is also very useful. Knowledge of AI planning and knowledge representation techniques is an advantage, but not strictly necessary. Contact:E: Dr Patrik HaslumBuilding a General Game PlayerProject Code: S_03Supervisor:Dr. Patrik Haslum, NICTA, CSLOutline:General Game Playing is an initiative from Michael Genesereth at Stanford university, which promotes building programs that are capable of playing arbitrary games by looking at the description of their rules. This differs from usual game-playing programs such as Deep Blue which are dedicated to one particular game (namely chess in this case), but are unable to play any other game even if the rules are merely variants of those the system was designed for. Another area of artificial intelligence, AI planning, has developed very efficient general tools to solve problems consisting in finding a sequence of actions (or moves) to reach a goal state. The most successful planners use search guided by heuristics that they automatically generate from the problem description. This is directly applicable to single-player games, such as Sokoban, the Freecell card game, or the old Rubik's Cube, but not to multi-player games. Goals of the project:The project consists in building a general game playing agent by extending a to-be-chosen AI planner and its automatic heuristic generation to solve a class of multi-player games of complete information, described in the Game Description Language (GDL). The player should be able to connect to the Stanford game manager, and will hopefully be able to enter the next General Game Playing Competition -- there is 10,000 USD prize to the winner! Prerequisites:This project will suit a student with an interest in artificial intelligence and a desire to learn about various aspects of games and planning, with good problem-solving skills and decent programming skills. Some knowledge of these topics and of elementary logic would be appreciated but is not necessary. Contact:E: Dr Patrik HaslumDiagnosis by SAT AlgorithmsProject Code: S_04Supervisors:Dr. Alban Grastien, DPO Group (ANU) & Managing Complexity (NICTA) Dr. Anbulagan, DPO Group (ANU) & Managing Complexity (NICTA)Outline:Recent works on diagnosis of discrete-event systems propose the use of satisfiability algorithms for efficient computation. Diagnosis is determining what happens on a system (car, plane, telecommunication network, water supply network, etc.) from observations on this system. This is an important task for monitoring and maintenance of expensive or critical systems. The main difficulty is usually to manage complexity. Discrete-event systems is a modeling of dynamic systems based on discrete (i.e. not continuous) variables. Satisfiability is the problem of determining if the variables of a given Boolean formula can be assigned in such a way as to make the specified formula evaluate to TRUE. This 50-years-old problem is still very active and attracts a lot of attention as many problems in Computer Science can be translated into SAT. The possibilities of SAT solvers for diagnosis have not completely been investigated. Extensions of the recent results in order to improve both efficiency and expressiveness of this approach include:
References:A. Grastien, Anbulagan, J. Rintanen and E. Kelareva, Diagnosis of discrete-event systems using satisfiability algorithms, Proceedings of the 22nd AAAI Conference on Artificial Intelligence (AAAI-07), AAAI Press, 2007 J. Rintanen and A. Grastien, Diagnosability testing with satisfiability algorithms, in M. Veloso, ed., Proceedings of the 20th International Joint Conference on Artificial Intelligence, pages 532-537, AAAI Press, 2007.Contact:E: Dr Alban GrastienModeling Mail Servers for Diagnosis and DiagnosabilityProject Code: S_05Supervisor:Dr. Alban Grastien, DPO Group (ANU) & Managing Complexity (NICTA)Outline:A mail server is a system composed of several components interacting to send and receive e-mails. When a failure occurs (a mail was not received), it is very difficult to track the problem. The person responsible for maintaining the server has to search into Gigabytes of logs to determine what exactly happened. The task is very difficult as
Basically, the difficulty comes from the incremental implementation of the servers. This task requires time, expertise, and can be unsuccessful in some cases. We propose to use the diagnosis techniques to diagnose the server, and also to determine which level of verbosity is required for each software in the server (diagnosability). The system behaviour is described in a model and a reasoning is performed on the model to understand what happens on the system. You will be involved in the first task (modelling of a mail server) to enable the second task (diagnosability and diagnosis). The former is not independent from the latter and you will have to understand the diagnosis techniques and test your modelling to improve or refine the modelling. Prerequisites:This project requires knowledge in Linux and networks (and ideally, in mail servers). Knowledge in AI techniques is an advantage but not necessary. It suits particularly for students who want to have a practical application. Contact:E: Dr Alban GrastienArtificial Intelligence in Video GamesProject Code: S_06Supervisor:Dr. Adi Botea, DPO Group (ANU) & Managing Complexity (NICTA)Outline:Video games are an excellent testbed for artificial intelligence research. They model real-life features such as uncertainty and dynamic worlds, and are suitable for several AI areas such as planning, learning and heuristic search. Commercial games are a fast-growing multi-billion dollar industry. The contacts between industry and academia are more and more present, stimulated in part by the newly created AIIDE conference, meant to bridge the gaps between the two communities. We offer an opportunity to work on such an exciting application and possibly engage in a longer collaboration. Here are a few more concrete project ideas to begin with:
Prerequisites:We expect the applicants to have strong C++ programming skills and an interest in artificial intelligence. Links:
Contact:E: Dr Adi BoteaFactored PlanningProject Code: S_07Supervisors:Dr. Adi Botea and Dr. Sylvie Thiebaux, DPO Group (ANU) & Managing Complexity (NICTA)Outline:Planning is the problem of finding a sequence of actions that reach a goal starting from a given initial state. Factored planning is a relatively new family of decomposition approaches which are useful when a problem is too large to be solved as one piece and has an appropriate structure. We integrate this into the SuperCom project, where a large real-life system (e.g., the water network in a city) is monitored and automatically reconfigured to a normal state (using planning) when faults are detected. There are many possible factored planning methods, and many ways of improving their performance. The project will consist in helping in the design and implementation (in JAVA) of one such method. Prerequisites:We expect the applicants to have good programming and problem-solving skills and an interest in artificial intelligence. Joining this project is an excellent opportunity to learn about planning and do exciting and rewarding work that will be used to solve real problems. Opportunities exits for further collaboration, for instance as an honours or PhD student. Contact:E: Dr Adi BoteaImmobile Robots: Integrating Diagnosis and PlanningProject Code: S_08Supervisors:Dr. Adi Botea and Dr. Alban Grastien, DPO Group (ANU) & Managing Complexity (NICTA)Outline:An immobile robot (immobot) is a real-life system (e.g., a power grid) enhanced with AI capabilities for self control and self configuration. Diagnosis and planning are two techniques at the core of the AI component of an immobot. Diagnosis processes sensor data and makes inferences about the system status. When a fault is detected, planning tells how to automatically reconfigure the system back to a normal state. Goals of the project:The goal of the project is to design and implement a model that integrates planning and diagnosis.Prerequisites:The applicants should have good Java programming skills. The project can be extended into an honours and/or PhD program.Contact:E: Dr Adi BoteaE: Dr Alban Grastien How to Beat the World's Best Satisfiability SolversProject Code: S_10Supervisor:Dr. Jinbo Huang, DPO Group (ANU) & Managing Complexity (NICTA)Outline:Satisfiability (SAT) is the problem of determining whether a Boolean formula can evaluate to true (i.e., be satisfied) under some assignment of truth values to its variables. For example, the formula ((X or (not Y)) and (X or Y or (not Z)) and ((not X) or (not Y) or (not Z))) can be satisfied by assigning false to all three variables. Problems in many areas of AI and computer science can be reduced to SAT. Although SAT is NP-complete, problem instances arising from real-world applications often have special structure allowing them to be solved efficiently. Current SAT solvers are frequently able to solve instances with over a million variables. This project will involve a review of the latest SAT technology, particularly a class of algorithms known as clause learning, followed by an attempt to advance the state of the art by proposing new algorithms or improvements to existing algorithms. On the practical side, the student will have access to an existing simple (yet competitive) SAT solver, written in C++, and will modify it and conduct experiments to explore the possibilities of improving its performance. The final product could be either an extension of the existing solver or an entirely new solver, which can then be entered into next year's International SAT Competition. Contact:E: Dr Jinbo HuangKnowledge Compilation for Probabilistic PlanningProject Code: S_11Supervisor:Dr. Jinbo Huang, DPO Group (ANU) & Managing Complexity (NICTA)Outline:Probabilistic planning is concerned with domains where the agent wishes to achieve a goal with high probability when its initial state and/or action effects are probabilistic. It is possible to encode probability distributions, as well as the entire probabilistic planning domain, using propositional logic. One can then compile the logic into a special form that allows efficient inference. The key is that the compilation can often be very efficient as it exploits the structure of the planning domain. This compilation-based approach has greatly improved the efficiency and scalability of probabilistic reasoning with Bayesian networks, which has much in common with probabilistic planning. We would therefore like to investigate the possibility of carrying this success over to probabilistic planning. The main challenge would be to design and implement new domain encoding methods, and new algorithms to work with these encodings, that can solve a given type of planning task (e.g., conformant planning, contingent planning, etc.). The compilation part will be accomplished using an existing state-of-the-art knowledge compiler. The new planner implemented can then be entered into the next International Planning Competition. Contact:E: Dr Jinbo HuangConstraint-Based PlanningProject Code: S_12Supervisor:Dr. Jussi Rintanen, DPO Group (ANU) & Managing Complexity (NICTA)Outline:SAT and CSP are general framework for solving arbitrary problems that can be viewed as constraint satisfaction. Best algorithms for SAT/CSP are able to solve very challenging problems from application areas such as diagnosis, model-checking and planning. However, the use of generic SAT/CSP solvers has the drawback that the utilization of heuristics and reasoning techniques specific to application areas is very difficult. Therefore in some cases it is a more efficient to implement specialized constraint solvers for a problem representation specific to the application. Goals of the project:The goal of the project is to implement a very efficient planner based on SAT/CSP technology by integrating the constraint solver with the problem representation that is specific to planning problems. This way it is possible to utilize constraint reasoning techniques for planning more efficiently while still benefiting from powerful inferences sanctioned by techniques like constraint propagation and clause-learning. Contact:E: Dr Jussi RintanenDesign of floor layouts for residential and industrial buildings as constraint satisfactionProject Code: S_13Supervisor:Dr. Jussi Rintanen, DPO Group (ANU) & Managing Complexity (NICTA)Outline:An important problem in the design of residential and other buildings as well as manufacturing plants is the layout of rooms and equipment. The objective when constructing the building is to minimize its construction costs while still providing for the required functionalities. Typically the design of floorplans for residential and office buildings is the job of an architect. The layouts of industrial plants are similarly the task of engineers who plan the production. Goals of the project:The purpose of this project is to cast the layout problem as a constraint satisfaction problem and demonstrate its utility as a fast prototyping method and possibly as a method for finding non-obvious designs a human designer might overlook. Finding better designs is an important factor in cost effectiveness of a building. Constraint Satisfaction provides a general and powerful framework for expressing and solving design problems. Depending on the desired extent of the project, the work could initially restrict to the design of residential buildings. Constraints on the room layout express spatial relations between the locations of the rooms, between building features such as doors and windows, as well as typical pieces and setups of furniture. For example, a typical living room in a small apartment includes a couch, armchairs, a table and television, the placement of which has to satisfy certain constraints derived from their intended functions. Contact:E: Dr Jussi RintanenChronicle Construction from FSM ModelProject Code: S_14Supervisor:Dr. Alban Grastien, DPO Group (ANU) & Managing Complexity (NICTA)Outline:Diagnosis is determining what happens on a system (car, plane, telecommunication network, water supply network, etc.) from observations on this system. This is an important task for monitoring and maintenance of expensive or critical systems. Chronicles Recognition [Dousson, 2002] is a diagnosis technique that has been successfully applied to several domains including the diagnosis of telecommunication networks and the diagnosis of cardiac arrhytmia (see for instance [Fromont et al., 2003]). Basically, a chronicle is a list of temporally-constrainted events. A chronicle is recognised when an instance of each event is found in the flow of observations that satisfies the constraints. The chronicle are usually made ``by hand''.=20 Automatic generation of chronicles are challenging, although it has investigated for Petri-Nets-modelled system [Guerraz--Dousson, 2004] or cardiac arrythmias (through Machine Learning techniques). In this project, we propose to automatically generate chronicles-like patterns from a finite-state-machine-based model. This task is difficult as it must avoid ambiguities and false recognitions. The work will be theoretical but also applied to the monitoring of mail servers. References:Elisa Fromont, Marie-Odile Cordier, Ren=E9 Quiniou, and Alfredo Hernandez: Kardio and Calicot: a comparison of two cardiac arrhythmia classifiers , AIME'03 Workshop: Qualitative and Model-based Reasoning in Biomedicine Christophe Dousson : Extending and unifying chronicle representation with event counters, ECAI'02 Bruno Guerraz and Christophe Dousson Chronicles Construction Starting from the Fault Model of the System to Diagnose, DX'04 Contact:E: Dr. Alban GrastienPlanning in networksProject Code: S_15Supervisor:Dr. Jussi Rintanen, DPO Group (ANU) & Managing Complexity (NICTA)Outline:Many planning problems involve the transportation network, a power distribution network, a water distribution network or connections (a binary relation) between nodes in general. Existing planning languages poorly support network-oriented planning, which motivates the introduction of expressive and efficient languages for expressing and solving such planning problems. The project is an investigation of implementation techniques for network-oriented planning. The basis is a planning language which is based on a modal logic for expressing network properties. The goal is the development of an efficient planning system for this language by using heuristic search and the state space representation of the planning problem. The main goal of the project is an implemented planning system, but the topic provides ample possibilities for more theoretical/analytical excursions to more fundamental questions about network-oriented planning. Contact:E: Dr Jussi RintanenLogical Analysis of Constraint ModelsProject Code:L&C_01Supervisor:A/Prof Peter Baumgartner, NICTAOutline:This project is embedded in a bigger NICTA project, which develops a new software platform for solving large scale industrial combinatorial (optimisation) problems. For example, a toy problem is the "n-Queens problem": is it possible to put n queens on a chess board so that no one threatens any other? Now, problems often admit symmetric solutions, or some decision variables depend functionally on others. It would be beneficial to know about such and related properties. The summer scholar will work on this topic by testing general methods---automated theorem provers--- to analyse problem specifications for their properties as indicated above. This will be a research-oriented internship, and, ideally, the software to be developed will be incorporated into the overall project software. Contact:E: Dr Peter BaumgartnerW: Dr Peter Baumgartner An Automated Theorem Prover That can do ArithmeticProject Code:L&C_02Supervisor:A/Prof Peter Baumgartner, NICTAOutline:An automated theorem prover is a program that tries to automatically prove valid (or disprove) formulas given to it. Many applications of automated theorem proving require reasoning modulo some form of (integer) arithmetic, e.g. for software verification. Unfortunately, theory reasoning support for the integers in current theorem provers is too weak for practical purposes. This summer scholar project is about implementing a theorem prover that can do better. The theory behind it has already been developed (by and large). A nice outcome would be a prototypical implementation and running some experiments with it. There is ample room to refine the theory and bring in own ideas. Contact:E: Dr Peter BaumgartnerW: Dr Peter Baumgartner Natural Language Query AnsweringProject Code:SML_03Supervisor:Dr Scott Sanner, SML, NICTAOutline:Ever wanted to ask your computer a question in natural language and have it respond with the answer you wanted? It sounds like a daunting task, but many of the foundations of such a technology are well understood. It is simply the process of inferring all possible relevant answers from an Internet full of content that is intractable. In this project, we view this inference task as one of reinforcement learning (i.e. how to learn to make a relevant sequence of inferences). In the end, our goal is to build a small prototype of this technology for natural language query answering in Wikipedia. Contact:E: Dr Scott SannerFinding Latent Variables in Reinforcement LearningProject Code:SML_04Supervisor:Dr Scott Sanner, SML, NICTAOutline:Reinforcement learning is the task of how to learn to make optimal sequential decisions when you can only sample experience from acting in a (simulated) environment. One of the greatest obstacles to practical reinforcement learning is finding the right features to use. In this project, we will examine automated feature induction methods via latent variable learning. The resulting reinforcement learning agent can be applied to one of the following tasks:
Contact:E: Dr Scott SannerCombining Machine Learning and LogicProject Code:SML_05Supervisor:Dr Scott Sanner, SML, NICTAOutline:There is great potential to exploit the powerful abstraction capabilities of logic for use in machine learning tasks. In this project, you would choose a task (natural language information extraction, query answering, probabilistic database inference, etc...) and implement a small learning and inference engine based on an appropriate logic (first-order, higher-order, description, or modal logic, etc...) and the recently studied structured learning tool of conditional Markov random fields (CRFs). A large library of Java tools will provide most of the underlying code needed to implement this project. Following implementation, the student will compare their work to state-of-the-art machine learning tools. This project offers the student the chance to learn about logic and machine learning, and how they can be combined for practical applications. Contact:E: Dr Scott SannerMars Rovers and Traffic ControllersProject Code:SML_06Supervisor:Dr Scott Sanner, SML, NICTAOutline:Markov decision processes (MDPs) are a theoretical tool for modelling sequential decision making problems and their optimal solution. Recent advances in the theory of MDPs permit efficient solutions to problems with both continuous state and action spaces. Such models are highly appropriate for planning in both Mars Rovers and Traffic Controllers (just to name two examples). In this project, you would choose one of these problem domains (or perhaps another you can suggest) and implement an (approximately) optimal planning system for this task. This project offers the chance for the student to learn about the theory of optimal sequential decision making and its application to practical problems. Contact:E: Dr Scott SannerScheduling for a High-performance Computing ClusterProject Code:SML_07Supervisor:Dr Scott Sanner, SML, NICTAOutline:We have a partner at the ANU who is looking for better process schedulers for their supercomputing systems. This requires performing some data analysis on their server logs to develop a model of process performance (based on a number of observed features) and then feeding this data to a constrained scheduler (which we have in-house). This project has great potential to make a substantial impact in the world of high-performance computing. Contact:E: Dr Scott SannerKnowledge Discovery from Large Honeypot Data for Internet SecurityProject Code:SML_08Supervisor:Dr Warren Jin, InfoEng, NICTA VISTAOutline:The Internet has evolved into a platform for all kinds of security-sensitive services/applications. For example, online banking and payment have become part of today's way of life. Internet security has been becoming knottier as cyber attacks have significantly increased in volume, coordination and sophistication due to the wide spread of worms and botnets Rentable botnet services, e.g., have resulted in sophisticated botnets becoming an effective and popular tool for committing online crime in recent years. Honeypots, as information system traps, are monitoring or deflecting malicious attacks on the Internet. A honeypot is a computer or network trap dedicated to attract and monitor malicious attacks. The NICTA honeypot project have access a so-called SGNet honeypot database. More than 25 SGNet honeypots have been collecting malicious Internet traffic since Oct 2007. We currently explore data mining techniques, say, data visualisation and outlier detection, to discover knowledge for Internet threats from honeypot data for proactive security purposes. This summer scholar project will study and/or develop outlier detection techniques which could handle mixed data features, multiple data sources, large volume of data or rapidly-evolving data streams. Specifically, we will investigate online or scalable outlier detection techniques, which could be used to detect rare but interesting attack patterns from honeynet data. The project involves some data mining, computer network and database techniques. The summer scholar will be supervised by Dr Warren Jin on a day-to-day basis but will have the opportunity to collaborate with the other project members and to explore your own ideas related to the project. You will experience the work environment in both ANU and an industrial research organisation. You will have the opportunity to get hands-on experience with some of the latest technology, e.g., stream mining, outlier detection and honeypots. Prerequisites:Applicants are expected to have a major in computer engineering or mathematics, preferably with excellent programming skills (in Java and/or C/C++) for implementation. Contact:E: Dr Warren JinPeer-to-Peer Botnets Investigation based on Distributed HoneypotsProject Code:SML_09Supervisor:Dr Warren Jin, InfoEng, NICTA VISTAOutline:Cyber attacks have significantly increased in volume, coordination and sophistication due to the spread of worms and bots. A botnet is a set of compromised machines, i.e., bots, under a common Command-and-Control mechanism, often used for nefarious purposes. A bot behaves like a complex of worms, rootkits and Trojan horses. It could automatically exploit, propagate and provide a remote control channel to its master. Botmasters providing stealthy botnet services for rent have deteriorated Internet security in recent years. The appearance of P2P botnets, which communicate via self-adaptable peer-to-peer protocol, makes security experts more difficult to observe and prevent cyber attacks. Goals of the project:This summer scholar project aims to investigate p2p bots based on the wonderful SGNet honeypot database. For example,
The summer scholar will be supervised by Dr Warren Jin on a day-to-day basis but will have an opportunity to collaborate with the senior security experts from DSTO. You will experience the work environment in both ANU and an industrial research organisation. You will have the opportunity to get hands-on experience with some of the latest technology, e.g., botnets, honeypots and security defence. Prerequisites:Applicants are expected to have a major in computer engineering or science, preferably with excellent programming skills (in Java and/or C/C++) for implementation. Knowledge of data mining or experience with computer network is a plus. Contact:E: Dr Warren JinIntelligent Prediction for Spatial DataProject Code:SML_10Supervisor:Dr Warren Jin, InfoEng, NICTA VISTAOutline:Spatial data are ubiquitous nowadays, such as medium house price and population composition of suburbs, social-economic indexing for each postcodes, national rainfall and natural resource distribution. Conducting intelligent prediction, such as water resource forecasting or population projection, from multiple spatial data is a quite useful but challenging issue. Those spatial data, e.g., could have different accuracy, different resolution, different time periods and different reliability. Goals of the project:This summer scholar project aims to develop and implement an intelligent prediction model which could integrate various information sources. The model could be based on hierarchical Bayesian models, continuous time Bayesian network, Markov random fields, or hidden Markov model. The prediction might comply with some spatial constraints or domain knowledge. The project would involve various techniques including GIS data manipulation and display, spatial data mining and statistical data processing. Prerequisites:Applicants are expected to have a major in information technology, computer engineering or statistics, preferably with excellent programming abilities (MATLAB, C/C++ or JAVA) OR sound mathematical/data mining/GIS background. Supervision will be provided on a day-to-day basis by Dr. Warren Jin. Contact:E: Dr Warren JinOntology-Driven Text MiningProject Code:SML_11Supervisor:Dr Warren Jin, InfoEng, NICTA VISTAOutline:Text mining has wide applications, such as e-marketing and digital forensics. Digital forensics involves understanding specific aspects of digital evidence and the general forensic procedures used when analysing any form of digital evidence. Digital evidence can be any information of probative value that is either stored or transmitted in a binary form, such as Emails, Office documents, computer system log files, as well as digital audio and video. It can be used to decide whether a crime has been committed and can provide a link between a crime and its victim or a crime and its perpetrator. Goals of the project:This summer scholar project aims at further developing or implementing effective text mining techniques to analyse textual information, especially large Email sets like Enron emails. Ontology can bring expressive background knowledge, e.g., the query “mouse - computer” is semantically looking for something related with animals rather than electronic device. Driven by ontology, textual information then can be indexed, summarised or analysed on conceptual level. In addition, this project will take into account of other specific characteristics of textual information, such as metadata or contextual data of an email. Prerequisites:The project requires good programming skills in Java. Some familiarity with text mining is also useful. Besides experiencing the work environment in both ANU and NICTA, the scholar will learn to handle real-world problems using new techniques. Supervision will be provided on a day-to-day basis by Dr. Warren Jin. Contact:E: Dr Warren JinApply Dynamical Bayesian Network to Query Digital ForensicsProject Code:SML_12Supervisor:Dr Nianjun Liu, InfoEng, NICTA VISTAOutline:Digital forensics undertakes the post-mortem reconstruction of the causal sequence of events arising from an intrusion perpetrated by one or more external agents, or as a result of unauthorised activities generated by authorised users, in one or more digital systems. The field of digital forensics covers a broad set of applications, uses a variety of evidence and is supported by a number of techniques. Application areas include forensic accounting, law enforcement, commodity flow analysis and threat analysis. Forensic investigations often focus on unusual and interesting events that may not have arisen previously. A major objective of a digital investigation is to extract these interesting pieces of evidence and to identify the causal relationship between this evidence. Goals of the project:This project aims at extending an existing Dynamical Bayesian Network model developed for digital forensics by investigating a number of possible topics. The model developed uses a Bayesian network and hidden Markov model network structure to (i) estimate typical digital crime scenario models from data and (ii) given such models, infer the most likely criminal act given current observations and past criminal acts. First, the addition of multi forward linkages in the BN+HMM network structure to allow direct linkages between BN nodes at consecutive time intervals. Second is the design of an SQL based query tool to explore the activities of criminals and their interactions and explain what happened in the past, as well as predict what will happen in the near future. Finally, the application of a graphical model to data mining of relational digital forensic databases, including construction of a relational pattern structural database for known types of digital crime portfolios and their associated forensics Bayesian Network models. Prerequisites:Applicants must have a major in information technology, computer science, or electrical engineering, preferably with excellent programming abilities (MATLAB, C/C++ and JAVA) OR strong mathematical/machine learning/data mining/statistics background. Contact:E: Dr Nianjun LiuUniversal Artificial IntelligenceProject Code:AI_01Supervisor:Dr Marcus Hutter, RSISE@ANU and SML@NICTA, CanberraOutline:The dream of creating artificial devices that reach or outperform human intelligence is an old one. Most AI research is bottom-up, extending existing ideas and algorithms beyond their limited domain of applicability. The information-theoretic top-down approach pursued in [Hut05] justifies, formalizes, investigates, and approximates the core of intelligence: the ability to succeed in a wide range of environments [LH07]. All other properties are emergent. Goals of this projectThe fundamentals of UAI are already laid out, but there are literally hundreds of open questions (see the exercises in [Hut05]) in this approach that have not yet been answered. The complexity ranges from suitable-for-short-projects to full PhD theses and beyond. Requirements
Student's gain
Literature
Contact:E: Dr Marcus HutterW: Dr Marcus Hutter Human Knowledge Compression ContestProject Code:AI_02Supervisor:Dr Marcus Hutter, RSISE@ANU and SML@NICTA, CanberraOutline:Being able to compress well is closely related to intelligence as explained below. While intelligence is a slippery concept, file sizes are hard numbers. Wikipedia is an extensive snapshot of Human Knowledge. If you can compress the first 100MB of Wikipedia better than your predecessors, your (de)compressor likely has to be smart(er). The intention of the Human Knowledge Compression Prize [Hut06] is to encourage development of intelligent compressors/programs. Goals of this projectSome of the following four subgoals shall be addressed:
Requirements
Student's gain
Literature
Contact:E: Dr Marcus HutterW: Dr Marcus Hutter On the Foundations of Inductive ReasoningProject Code:AI_03Supervisor:Dr Marcus Hutter, RSISE@ANU and SML@NICTA, CanberraOutline:Humans and many other intelligent systems (have to) learn from experience, build models of the environment from the acquired knowledge, and use these models for prediction. In philosophy this is called inductive inference, in statistics it is called estimation and prediction, and in computer science it is addressed by machine learning. The problem of how we (should) do inductive inference is of utmost importance in science and beyond. There are many apparently open problems regarding induction, the confirmation problem (Black raven paradox), the zero p(oste)rior problem, reparametrization invariance, the old-evidence and updating problems, to mention just a few. Solomonoff's theory of universal induction based on Occam's and Epicurus' principles, Bayesian probability theory, and Turing's universal machine [Hut05], presents a theoretical solution [Hut07]. Goals of this project
Requirements
Student's gain
Literature
Contact:E: Dr Marcus HutterW: Dr Marcus Hutter Universal Induction versus No Free LunchProject Code:AI_04Supervisor:Dr Marcus Hutter, RSISE@ANU and SML@NICTA, CanberraOutline:Solomonoff's theory based on Occam's razor provides a rigorous and formal method for inductive inference, prediction, and time series forecasting, by essentially uniquely specifying a universal model class and prior. This model seems to solve the long outstanding (philosophical) problem of induction and many other deep statistical questions [Hut07]. On the other hand, the No-Free-Lunch theorem(s) [WM97] state that all optimization or search algorithms are on average equally good/bad, if a uniform average over the space of all functions is taken. So NFL believers conclude that we need a prior, biased to our particular problem class at hand, that is, there is no universal (problem independent) solution to the induction problem. There is an ongoing battle between believers in Occam's razor and believers in no-free-lunches [Sto01,SH02]. Goals of this project
Requirements
Student's gain
Literature
Contact:E: Dr Marcus HutterW: Dr Marcus Hutter Hash-consing For Efficiency in the HOL4 KernelProject Code: L&C_03Supervisor:Dr. Michael Norrish, L&C Program, NICTAOutline:The heart of the HOL4 theorem-proving system is its kernel, where the fundamental types of terms and theorems are implemented. Because all theorem-proving activity ultimately devolves to this level, it is clearly critical that the implementations be as efficient as possible. This project is about investigating whether one possible optimisation is worthwhile. The optimisation in question is known as "hash-consing", and involves setting the kernel up so that there is only ever one copy of any given value in memory. This makes comparison a constant-time operation (as opposed to a tree traversal), and also allows for efficient implementation of the other operations. This project would suit someone who is keen to program on a real system (HOL has been under development around the world since the mid 1980s), who is interested in data structures and algorithms (particularly over trees!), and who is happy to program in the SML programming language. Supervision will be provided on a day-to-day basis by a senior researcher. Contact:E: Michael.NorrishUseful GamesProject Code: AI_05Supervisor:Dr. Jochen Renz (ANU)Outline:Games are becoming closer and closer to real life and some tasks we have to solve in games are similar to tasks we have to solve in our real life. One surprising difference is that for many people solving the tasks in games is much more fun and less tiring than in real life. Consequently, it can be useful to transform tasks or work we do not like into a game and do our work by playing a game--or even better, let other players do the work for us. Goals of project:In this project we will analyse different possible tasks, select one of them, transform it into a game and test its performance. Prerequisites:Interested students should be interested in playing games and have experience with different game concepts. Supervision will be provided on a day-to-day basis by Dr. Jochen Renz. Contact:E: Jochen RenzQualitative Spatial Representation and ReasoningProject Code: AI_06Supervisor:Dr. Jochen Renz (ANU)Outline:In computer science and engineering, spatial information (such as the spatial description of a room or a city) is often represented using coordinate systems, but this is not the way humans deal with spatial information. We often describe spatial features by specifying the relationship between objects in space (for example, the book is "on" the table, "to the left of" the screen). This is what is done in the area of Qualitative Spatial Representation and Reasoning, an important subfield of Artificial Intelligence. It is possible to define many different spatial relations for different aspects of space such as direction relations, distance relations, size relations etc. and to express spatial knowledge using these relations. We offer different projects within this framework. Here is a small selection, other topics available on request:
Since reasoning with these relations is a very complex problem, it requires a detailed theoretical analysis of the different relations in order to find efficient reasoning algorithms for different sets of relations. However, recent advances in the area allow to make this theoretical analysis and to find efficient reasoning algorithms automatically. In this project we look at different spatial calculi, test if we can find efficient algorithms automatically and try to find out what makes reasoning hard.
Using qualitative spatial relations is very helpful for describing routes and for providing navigation assistance as such expressions are easier for users to understand and to remember. In this project we will try to extract qualitative spatial relations using maps, landmarks and the current location of users, plan routes according to different criteria and express the routes in terms of qualitative spatial descriptions.
There are several ways of representing biological entities like proteins, molecules or cells, which are often too complex for applying data-mining, machine learning or other computational techniques. Many of the functional properties of biological entities and of their interactions depend on spatial properties, e.g. a protein might only interact with another protein if it is folded in a way that some of its parts are accessible from the outside to other proteins. Goals of project:The idea of this project is to find examples of biological properties and interactions where spatial relationships are important, identify and categorise these relationships, and based on this find simple functional representations of biological entities which could be used for applying computational techniques. Prerequisites:A background in artificial intelligence or biology/bioinformatics would be desirable but not essential. Supervision will be provided on a day-to-day basis by Dr. Jochen Renz Contact:E: Jochen RenzUser-friendly navigation with Google MapsProject Code: AI_07Supervisor:Dr. Jochen Renz (ANU)Outline:GPS devices for identifying the location of users are becoming more widespread and are now available in many mobile phones and PDA's. Using these devices it is easily possible to assist the user in finding the best route to a given goal with respect to certain optimisation criteria, usually the shortest or fastest route. In this project we look at more user-friendly optimisation criteria such as finding the easiest route between two locations. We will use the Google Maps API in order to obtain the necessary information for the route optimisation algorithms we develop in this project and also in order to visualise the routes. Supervision will be provided on a day-to-day basis by Dr. Jochen Renz Contact:E: Jochen RenzTrust and ReputationProject Code: T&S_01Supervisor:Dr. Jochen Renz (ANU)Outline:In online transactions we often deal with people we do not know. How can we possibly be sure about who we can trust and who we should better not trust. A popular method is to use a reputation system which keeps track of the past performance of every user of the system and which assists users in forming their opinion about whether another user can be trusted. Even though many reputation systems are being used (the most prominent is probably the eBay Feedback System), their performance is not satisfactory at all and many fraudulent transactions are still happening. Goals of project:The aim of the project is to develop and implement a reputation system and to compare its performance with existing systems. Prerequisites:Interested students should have experience with online transactions and existing reputation systems such as eBay. Supervision will be provided on a day-to-day basis by Dr. Jochen Renz Contact:E: Jochen RenzAutomatic Prediction of Scientific DiscoveriesProject Code: AI_08Supervisor:Dr. Tiberio Caetano, CSL, NICTA SMLOutline:In this project you will have the opportunity to investigate questions of the following type:
The main idea will be to get hold of a substantial amount of data on the scientific literature - namely all scientific articles published in the top 5000 journals in all fields since 1945 - and to analyze the citation patterns of these articles mathematically using computers. Every article cites a number of other articles and this in fact establishes a big directed acyclic graph that is a proxy for the evolution of science in a period. We just need to mine this data intelligently, using sophisticated mathematical tools that you will learn during the course of the project. These include Graphical Models, Markov Random Fields, Topic Models, Dirichlet Processes, Pitman-Yor processes and lots of other fun and useful maths. Goals of project:
Prerequisites:Good programming skills in C or C++. Good maths background. Willingness to change the the landscape of knowledge.Reference:[1] W: sciencemagE: Dr Tiberio Caetano Modelling Human Encounter NetworksProject Code: AI_09Supervisors:Dr. Tiberio Caetano, CSL, NICTA SMLDr Sebastien Ardon, NICTA Outline:Human activity leads most people to come in close proximity of some other people on a daily basis, forming encounter networks. Sources of human encounter range from social activity, to random encounters in public transports. While seemingly random, a large proportion of encounters are the result of social activities with highly structured inter-dependencies both in time and space. Encounter networks aim at capturing both the spatial and temporal properties of human encounters. While the dynamics of encounter networks are likely to be different for specific socio-economic, and/or cultural groups, the fact is that they remain sparsely studied due the lack of available data. Applications in the computer networking and epidemiology fields include the evaluation of opportunistic networking communication, virus/disease propagation in urban environment, evaluation of gossip-based information dissemination algorithms. The project aim is to develop and validate a model for human encounter networks and to explore sampling methods for this model, in the view to drive a computer simulation tool. A validation method should be outlined and implemented, allowing experimenters to check for the validity of the output. Goals of the project
All these steps will be carried out under close supervision of Drs Ardon and Caetano. Reference:
Properties of n-type ribbon-grown silicon for solar cellsProject Code: DE_02Supervisor:Dr Daniel MacdonaldOutline:Ribbon-grown silicon is a low cost alternative to traditional silicon wafers for making solar cells. The silicon is grown in long, thin ribbons, eliminating the need to saw wafers from a large ingot. This reduces processing costs and silicon wastage through saw-dust. However, ribbon-grown wafers contain more defects and impurities, which may lead to lower cell efficiency. Usually, ribbon silicon is p-type. However, there are some reasons to believe that n-type ribbons may be more tolerant of certain important impurities. This project will study the electronic properties of n-type ribbon-grown silicon wafers, provided by a large international manufacturer, to see if they are suitable for solar cell production. Goals of the project:
Prerequisites:A reasonable knowledge of semiconductor physics and devices.Contact:E: Dr Daniel MacDonaldAssessing cheap silicon for solar cellsProject Code:Solar_01Supervisors:Dr Keith McIntosh, and Dr Dan MacDonaldOutline:Most solar cells are fabricated from silicon. The silicon is crystalline and relatively pure, which makes it expensive to produce, but some manufacturers are reducing its cost by crystallising the silicon more rapidly and by using less pure reactants. One problem faced by these manufacturers is how to assess their silicon. Usually, silicon is assessed by measuring its conductance under illumination to calculate the "lifetime" of its electrons. The cheap silicon prevents this measurement at room temperature because crystal defects "trap" electrons preventing them from contributing to the silicon's conductance. At the ANU, we have recently developed an instrument to circumvent the inhibitive "trapping" by performing the measurements at high temperatures. The project involves the assessment of cheap silicon over a range of temperatures. It would suit a student who enjoys experimentation and physics. The student would learn theory relevant to solar cells and semiconductors, and would gain experience in experimental physics. Contact:E: Dr Keith McIntoshE: Dr Dan MacDonald
Research in solar thermalAt the ANU Solar Thermal Group, we’re working on a next-generation solar thermal energy system including a new design for large dish concentrators (following our previous Big Dish design) and some new work on designing ammonia-based solar chemical energy storage. We’re working on different aspects of this, including structural design, chemical process design, fluid dynamics, system modelling, control, optical analysis and characterisation and more. As far as summer research goes, we haven’t got a fixed project scope in mind, but if you would like to work in this area with us, particularly if you are considering a final year project in Engineering at ANU, then contact us, and we can talk about potential projects. For more information about what’s happening in solar thermal energy at ANU, see these sites:
Community Detection in Complex NetworksProject Code: CN_1Supervisors:Professor David Hill, Federation Fellow, InfoEngMr Lachlan Black-hall, PhD student, InfoEng Outline:Complex networks are used to represent large and complex systems like the electricity network, aircraft or robots working cooperatively and even the computers on the internet. There is significant interest in determining what subgroups (or communities) exist within these networks and how they interact to explain the action of the network as a whole (i.e. How one power station affects the rest of the electricity network). Methods of identifying and classifying these communities will lead to fundamental advances in preventing blackouts in electricity networks, increasing aircraft safety in air traffic control and finding better and faster ways of transmitting data through the internet. Goals of the project:Through this project we are aiming to develop simple, efficient and computationally tractable ways of determining the location and membership of different communities within complex networks. Students will gain an insight into PhD level research in a variety of different fields including control theory, system and parameter estimation, graph theory and statistics as well as having fun solving a current major research question. Prerequisites:Students will require basic systems theory and statistics. An ability to program in MATLAB will be helpful but not required. Contact:E: Prof. David HillE: Mr Lachlan Black-hall Network Motifs in Complex NetworksProject Code: CN_2Supervisors:Professor David Hill, Federation Fellow, InfoEngMr Lachlan Black-hall, PhD student, InfoEng Outline:Complex networks are used to represent large and complex systems like the electricity network, aircraft or robots working cooperatively and even the computers on the internet. We are very interested in determining the fundamental motifs (building blocks) that can be used to represent these networks. An understanding of these basic building blocks (motifs) will allow us to create better ways of designing and analysing complex networks in the future. This will lead to fundamental advances in the design of fault tolerant electricity networks as well as new insights into the air traffic control problem of getting large numbers of aircraft safely landed at airports. Goals of the project:Through this project we are aiming to develop a catalogue of motifs that can be used to represent an arbitrary complex network. Furthermore we aim to determine the relationship between these motifs so that we can design and analyse networks at a higher level of abstraction, thus drastically simplifying the problems currently faced. Students will gain an insight into PhD level research in a variety of different fields including control theory, system and parameter estimation and graph theory as well as having fun solving a current major research question. Prerequisites:Students will require basic systems theory and some university level calculus and algebra. Contact:E: Prof. David HillE: Mr Lachlan Black-hall Bounded Look-Ahead Techniques for Planning in Hybrid DomainsProject Code: S_01Supervisor:Dr. Patrik Haslum, DPO Group (ANU) & Managing Complexity (NICTA)Outline:Characteristic of planning systems is that they develop a complete plan of action for achieving a specified goal, before that plan is considered for execution. Reactive systems, on the other hand, look only at the present state in order to select the next action to execute. In between these extremes are methods using bounded look-ahead, that is, developing a plan for a limited number of actions or a limited time into the future, and using this plan to make the choice of the next action to execute. In control engineering, this method is known as "model predictive" or "receding horizon" control, and it has been successfully applied to hybrid control problems (involving mixed discrete and continuous states). Goals of the project:The aim of this project is to apply the method of model predictive control to planning problems, in particular problems with such hybrid domains (involving metric time and resources) and where optimization of the cost of the plan carried out is a main concern. One of the main differences between control and planning is that while the objective of automatic control is typically to keep the controlled system stable or near a reference point, the objective in planning is to drive the controlled system to a specified goal state. Thus, one of the the main challenges of this project is taking into account the estimated cost of eventually achieving to the goal while optimizing a plan for the near future, particularly in the context of hybrid constraint optimization methods (such as LP/SAT, MILP and similar). Prerequisites:This project requires a good understanding of "formal stuff" (such as hybrid state/transition systems, admissible heuristics, etc.), though it's not particularly mathematical in nature. Programming skills are also likely to be required, though for the heaviest part of the implementation (optimization of hybrid constraint systems) there are probably systems/libraries available that can be used. Contact:E: Dr Patrik HaslumFormation Search and RescueProject Code:SEaCS_05Supervisors:Dr. Baris Fidan, Systems Engineering and Complex Systems Program, NICTA and Mr. Iman ShamesOutline:One of the most important capabilities of a formation of unmanned aerial vehicles should be its ability to detect a target of interest, this target may be different from mission to mission, it can be a victim of natural disaster or a hostile listening post, in either of the cases the formation should be able to search an area of interest for the target completely and with minimum time. For different sensing capabilities , the formation has to move in different path with a certain shape to maximize the likelihood of detecting the target if one exists and minimize the actual time needed for doing the task. Goals of the project:The project aim is to find the best trajectory and shape for a formation for different sensing constraints. Prerequisites:For this project, creative students enthusiastic in working new designs with self-initiatives are sought. In addition, being familiar with concept of programming is necessary (Capability of programming in any programming language such as C++, Java or Matlab will be a plus) The students will be supervised by Dr. Baris Fidan and Mr. Iman Shames. In addition they will have some interactions with a number of other research groups and academic visitors. Contact:E: Dr Baris FidanE: Mr Iman Shames Getting the Big PictureProject Code:SEaCS_06Supervisors:Dr. Baris Fidan, Systems Engineering and Complex Systems Program, NICTA and Mr. Iman ShamesOutline:"Formation flying provides an alternative approach to flying large structures in space," said Dr. Fred Hadaegh, a senior research scientist and manager of the Distributed Spacecraft Technology Program of JPL. The idea is implemented in the Terrestrial Planet Finder project which will rely on formation flying for one of its two observatories. Five separate spacecraft will work together to function as a single huge telescope.They will create a telescope powerful enough to distinguish the faint light of small Earth like planets from the much brighter light of the stars they orbit. Goals of the project:The project aims are to study spacecraft models and coordinate these models in a simulation environment, and point these telescope in a coordination in order to get the big picture of a distant extra terrestrial object. Prerequisites:The project will involve mathematical design, analysis, and simulations. Hence the student is expected to have a reasonable level of maths background. More important than the math background is the ability of the students to think in a multi-disciplinary fashion. The students will be supervised by Dr. Baris Fidan and Mr. Iman Shames. In addition they will have some interactions with a number of other research groups and academic visitors. Contact:E: Dr Baris FidanE: Mr Iman Shames Smart Self-configuring Sensor NetworksProject Code:SEaCS_07Supervisors:Dr. Baris Fidan, Systems Engineering and Complex Systems Program, NICTA and Mr. Iman ShamesOutline:Imagine a sensor network built up from smart mobile sensing nodes that may change their position to accomplish a certain task, this may need a more concentration of sensors in an area of interest where certain phenomena are happening, it may be making the sensors as scattered as possible to have a wider coverage of the environment in order to look for a particular thing or just exposing the sensor more to the environment for measuring a variable. Furthermore they may need to change their configurations to compensate a device failure, or a communication breakdown. The project outcome may be used in a broader context, and even may be implemented in a test environment. Prerequisites:For this project, creative students enthusiastic in working new designs with self-initiatives are sought. In addition, being familiar with concept of programming is necessary (Capability of programming in any programming language such as C++, Java or Matlab will be a plus) The students will be supervised by Dr. Baris Fidan and Mr. Iman Shames. In addition they will have some interaction with Prof. Brian Anderson, and interactions with a number of other research groups and academic visitors will be arranged. Contact:E: Dr Baris FidanE: Mr Iman Shames Multi-Agent Systems Simulation SoftwareProject Code:SEaCS_08Supervisors:Dr. Baris Fidan, Systems Engineering and Complex Systems Program, NICTA and Mr. Iman ShamesOutline:The topic of autonomous multi-agent systems has gained a lot of attention in recent years; hundreds of researchers start looking at different aspects of this field, with applications involving various kinds of sensor networks as well as teams of robots, unmanned aerial vehicles, and submarines. But, in the end of the day they want to show their method is working. In order to show that, guess what… they need to run programs and simulations! The problem they are dealing with is that usually their simulation software need extensive modifications each time another scenario is being tested. Now imagine a world with a systematic software capable of providing these diverse simulation scenarios with only changing some environmental variables in user-friendly interfaces… Mass simulations are just a few numbers and clicks away with the outcome of this project. The project outcome will be used excessively in the SWARM research group and is expected to be interest of many relevant research groups around the world as well. At a later stage, commercialisation of the software will be considered. Prerequisites:The interested students should be willing to familiarize themselves with different models for multi-agent systems and problem environments. Furthermore they should have a strong background and enthusiasm in programming and generating graphical user interfaces. The students will be supervised by Dr. Baris Fidan and Mr. Iman Shames. In addition they will have some interaction with Prof. Brian Anderson, and interactions with a number of other research groups and academic visitors will be arranged. Contact:E: Dr Baris FidanE: Mr Iman Shames Formation Tracking SystemProject Code:SEaCS_09Supervisors:Dr. Baris Fidan, Systems Engineering and Complex Systems Program, NICTA and Mr. Iman ShamesOutline:A formation of unmanned aerial vehicles (UAVs) is flying around in an environment; they may be hostile agents or friendly drones that needs to be tracked. The explanation for the need of tracking the former is pretty obvious for the latter the need may raise from a failure in the systems, e.g. loss of global positioning system (GPS). Tracking flying formation is an interesting topic, which requires integration of signal processing, control systems, and filtering, e.g. Kalman filters. Goals of the project:The project outcome is expected to be used as a corner stone for more complicated tracking algorithms capable of providing fast and reliable results in hostile and unknown environment. Prerequisites:The project will involve mathematical design, analysis, and simulations. Hence the student is expected to have a reasonable level of background in maths and programming (in at least one of C++, Java or Matlab). The students will be supervised by Dr. Baris Fidan and Mr. Iman Shames. In addition they will have some interaction with Prof. Brian Anderson, and interactions with a number of other research groups and academic visitors will be arranged. Contact:E: Dr Baris FidanE: Mr Iman Shames Hypersonic Flight ControlProject Code:SEaCS_10Supervisor:Dr. Baris Fidan, Systems Engineering and Complex Systems Program, NICTAOutline:Hypersonic flight (flight with speeds over Mach 5, i.e. 5 times the speed of sound) of aerospace vehicles in the atmosphere as well as for earth-to-orbit and atmosphere-re-entry missions has been of interest for the major space and aviation organisations of the world such as NASA for a long time. A particular interest is design of air-breathing hypersonic flight vehicles, i.e. air-vehicles taking the oxygen needed for combustion from air rather than carrying it on board in a rocket tube. The non-standard dynamic characteristics of air-breathing hypersonic flight vehicles together with various aerodynamic effects of hypersonic flight make the flight control of such vehicles challenging and interesting. Goals of the project:The aim of this project is providing new approaches and designs to the interesting problem above. The project outcome is expected to be used in some relevant projects pursued by a number of focused research groups and wider research organisations, including the Australian Hypersonic Initiative (AHI). Prerequisites:The project will involve mathematical design, analysis, and simulations of flight dynamic systems. Hence the student is required to know basics of flight/aircraft dynamics/mechanics and/or aerodynamics. Basic programming skills and familiarity with Matlab or Simulink are also needed. The students will be supervised Dr. Baris Fidan. Collaboration with the AHI and/or the UNSW@ADFA is expected. Contact:E: Dr Baris FidanSwitching or blending? - A comparative study on multiple model based observer designProject Code:SEaCS_11Supervisors:Dr. Weitian Chen, ANU, Prof. Brian Anderson, ANUOutline:Observer design aims to provide state estimation for dynamic systems based on limited information (measurements). For systems whose state variables are not available, observers are often required to be designed for the purpose of solving control problems. Therefore, observer design is an important research area in systems and control. Although it is solved well for an important class of systems called linear time invariant systems, observer design has experienced many difficulties in the course of moving toward complex systems such as time-varying systems and nonlinear systems. Multiple model based approach has a great potential to deal with those difficulties encountered in the observer design. Two different multiple model based methodologies have been proposed in the literature. One switches amongst observers designed for all models, while the other blends all those observers designed to form an overall observer. The question now is: which is better: switching or blending? Goals of the project:This project is to carry out a comparative study on the above mentioned observer design strategies through extensive simulations on different complex systems such as time-varying systems, nonlinear systems, and switched control systems. Those simulations will be carefully designed to reveal the advantages and disadvantages of each method. Furthermore, the possibility of new observer design will be explored. Prerequisites:Interested students should have some knowledge of observer design and be familiar with programming using Matlab. E: Dr Weitian Chen
Algorithms for Finding Complete Bipartite SubgraphsProject Code: S_09Supervisor:Dr. Jussi Rintanen, DPO Group (ANU) & Managing Complexity (NICTA)Outline:Many applications require the representation of data or knowledge in the form of a binary relation or a graph, like "X knows Y" or "location X has a road connection to location Y". If the number of nodes in a graph is high, like tens of thousands or millions, and the graph is dense, then representing all the edges explicitly one at a time leads to very inefficient processing of the data. By recognizing regularities in the graph and utilizing the regularity to represent the graph more compactly, big improvements in efficiency can be obtained. One regularity in many big and dense graphs is the presence of complete bipartite subgraphs N,M in which there is an edge between every node in N and every node in M. Each such subgraph could be represented simply by representing N and M, without enumerating all the |N| X |M| edges explicitly. Another regularity, complete subgraphs (cliques), can be compactly represented in terms of bipartite subgraphs too. Goals of the project:The goal of the research is to develop efficient algorithms for compressing the representation of graphs by identifying complete bipartite subgraphs in them. Finding the maximal bipartite subgraph of a graph is an NP-hard problem. Some polynomial time approximation algorithms with approximation guarantees exist, but it seems that they are not very practical for graphs consisting of tens of thousands or millions of nodes. The work will proceed by implementing and comparing existing approximation algorithms with and without approximation guarantees, and developing improvements to the existing algorithms and identifying heuristics for improving them. Contact:E: Dr Jussi RintanenA new algorithm to compute saddle points in differential gamesProject Code:CS_01Supervisors:Principal supervisor: Mr. Alex Feng, ANU and NICTA. Second supervisor: Dr. Weitian Chen, ANU Third supervisor: Prof. Brian Anderson, ANU and NICTAOutline:Game theory is a branch of applied mathematics that is used in the social sciences, biology, political science, and computer science. Game theory attempts to mathematically capture behaviour in strategic situations, in which an individual's success in making choices depends on the choices of others. The term "Differential Games" is applied to a group of problems in applied mathematics that share certain characteristics related to the modelling of conflict. In a basic differential game there are two players -- a pursuer and an evader -- with conflicting goals. The pursuer wishes, in some sense, to catch the evader, while the evader's mission is to prevent this capture. Differential games are closely linked to control systems. For a given control system, different inputs can be regarded as different players in a differential game. In such a situation, if each player has access to the state information of the control system, then the strategies these players take can be regarded as feedback control laws in the control system. In a differential game, a saddle point describes the situation when all players have tried their best to achieve their goals. Saddle point is an important concept in differential games and computation of saddle points is always of interest and challenge. Goals of project:This project aims to develop a new algorithm to compute saddle points of differential games. In this project, we consider the multi-player differential games and the research will focus on the following points:
Prerequisites:Mathematical modelling and proving are important in this project, and design of simple numerical examples will be needed. Interested students should be familiar of partial differential equations, ODEs, control systems, game theory and capable of programming in C/C++ or Matlab. Contact:E: Mr Alex FengA new algorithm to solve periodic Riccati equations in periodic control systemsProject Code:CS_02Supervisors:Principal supervisor: Mr. Alex Feng, ANU and NICTA. Second supervisor: Dr. Weitian Chen, ANU Third supervisor: Prof. Brian Anderson, ANU and NICTAOutline:Periodic control systems have many applications in the real world such as satellite control. Recently, periodic control systems have attracted much attention among control researchers. A frequently used control method in linear periodic systems is designing periodic controllers. To obtain these periodic controllers, one needs to solve periodic Lyapunov differential equations and periodic Riccati differential equations. Some certain algorithms to compute solutions of periodic Lyapunov differential equations and periodic Riccati differential equations have been developed. The development of new computational methods in solving periodic Riccati differential equations is both interesting and challenging. Goals of project:This project aims to develop software to solve periodic Riccati differential equations arising in linear periodic control systems using ideas recently developed by the proposed supervisors, to compare the software with existing software, and to compare simulation results with theoretical results. The student will normally develop software in Matlab or/and C/C++ and validate theoretical results. Prerequisites:Interested students should be capable of programming in Matlab and C/C++ and familiar with control systems, ODEs and computational complexity. Contact:E: Mr Alex FengThe world's fastest travelling salesmanProject Code:OR_01Supervisor:Dr Philip KilbyOutline:In the classic Travelling Salesman Problem (TSP) we are given a set of cities, and the distances between them. We want to find the order to visit the cities so that the distance is minimised. This is a very well-studied problem. In a variant of the TSP, we have lots of cities, but we only want to visit a subset of them. One way to solve this problem is to find a "grand tour" around all of the cities. When we are given the subset we wish to visit, we can generate the minitour in record-breaking time:simply visit them in the order they are visited in the grand tour. But building a grand tour to be used in this way is different to solving a standard TSP. We need to construct the grand tour quickly, but carefully. If we can do this, we will be able to make good, fast solutions to the TSP available in a bunch of new areas. This project will look at ways we can generate grand tours with nice properties. We will look at how we can interface with Geographical Information Systems to provide good routes easily. Goals of the project:
Prerequisites:
Contact:E: Dr Philip KilbyW: Dr Philip Kilby Using Mathematics to Match PeopleProject Code:A&D_01Supervisor:Dr Tiberio Caetano, CSL, NICTA SMLOutline:Think of the way people find partners these days. They go to parties, are attracted by appearance, start dating and after some time trouble begins. This is a routine. Since the "appearance score" of a person is immediately and reliably estimated without even the need of major physical proximity, people usually start relationships based on that, and usually finish relationships for other reasons. It is not acceptable that, being in the 21st century, we still use such inefficient methods to find matches. Since now we have information technology, we should use it in a smart way in order to create recommendations of which places you should go and when, and maybe even which particular people you should start talking to. In this project you will analyze real data for internet dating sites and develop and implement algorithms for matching people so as to optimize a global score function that has to do with the "aggregate happiness" of the collective matches. Each person has filled a detailed questionnaire with hundreds of questions and that information will be used to match people. The key novel research ingredient in this project however will be the fact that we will be using information about outcomes of previous relationships in order to optimize the "matching function" itself. This is technically a hard research problem which however can be tackled with modern mathematical and statistical tools. Goals of project:
Prerequisites:Good maths background and willingness to work on challenging new maths. Good C or C++ skills. E: Dr Tiberio Caetano
Particle Simulation using a FPGAProject Code: CS_04Supervisor:Dr. Dr Eric McCreathOutline:FPGAs have recently been the subject of interest for improving the performance of molecular dynamics simulations. FPGAs provide a very flexible way of pipelining the required computation, they have shown the ability of accelerating simulation performance. This project involves the development of a simple particle simulation on an FPGA. The simulation would be similar to the one developed by previous students on the Cell and a GPU. The simulation involves Lennard-Jones interactions and a simple O(n^2) algorithm(where n is the number of particles). Contact:E: Dr Eric McCreathCalculating optical flow using a FPGAProject Code: CS_05Supervisor:Dr. Dr Eric McCreathOutline:Vision processing is computationally expensive and this translates into : weight, power, and size, for a system doing the processing. When vision processing is required in autonomous or semi-autonomous robotic vehicles this becomes a significant issue. FPGA's are able to embed designs that can process a large amount of image data quickly, also their footprint is small in terms of : weight, power, and size. These factors make them ideally suited to vision processing in robotic vehicles. This project extends a previous student's work who partly implemented an optical flow algorithm within a FPGA. This project will involve completing this work with an aim to attach a camera and streaming to data directly from the camera into the FPGA(currently done off-line via a USB cable). This project would involve learning how to program an FPGA and control a camera. The project would be in collaboration with A/Prof Robert Mahony from the Department of Engineering and would be part of the 'Image based teleoperation of semi-autonomous robotic vehicles' ARC Discovery Grant he was awarded. Contact:E: Dr Eric McCreathHash-consing For Efficiency in the HOL4 KernelProject Code: L&C_03Supervisor:Dr. Michael Norrish, L&C Program, NICTAOutline:The heart of the HOL4 theorem-proving system is its kernel, where the fundamental types of terms and theorems are implemented. Because all theorem-proving activity ultimately devolves to this level, it is clearly critical that the implementations be as efficient as possible. This project is about investigating whether one possible optimisation is worthwhile. The optimisation in question is known as "hash-consing", and involves setting the kernel up so that there is only ever one copy of any given value in memory. This makes comparison a constant-time operation (as opposed to a tree traversal), and also allows for efficient implementation of the other operations. This project would suit someone who is keen to program on a real system (HOL has been under development around the world since the mid 1980s), who is interested in data structures and algorithms (particularly over trees!), and who is happy to program in the SML programming language. Supervision will be provided on a day-to-day basis by a senior researcher. Contact:E: Michael.Norrish
Particle Simulation on the Cell Broadband - Extending the types of interactionsProject Code: CS_03Supervisor:Dr. Dr Eric McCreathOutline:This project involves extending a simple molecular dynamics simulation that has been implemented on the Cell Broadband Engine to include some more complex particle effects. These effects would be chemical bonds and charge effects. This would enable the simulation to provide a basic molecular dynamics simulation of water, which is a most interesting and important molecule. This Particle Simulation project involves collaboration with A/Prof Alistair Rendell and aims to improve the performance and capability of high performance computer systems that undertake Molecular Dynamics. Molecular Dynamic simulations are key tools for molecular biologists for advancing science and medicine. This projects aim to improve the computing foundations that will improve the systems of the future. This projects involves the Cell Broadband Engine. This is a high performance processor which includes 8 SIMD processors, called SPE's, which are able to perform a large number of floating point operation very quickly (peak ~25 GFlops in each SPE). These SPEs can be programmed in c, as instructions do not have direct access to main memory programming them can be a little tricky! The project will involve learning how to program the SPEs. Note that the Cell is used within Sony's PS3 gaming machine. This is a very powerful and interesting architecture to program. The project would build on a system that was developed by a previous student that simply evaluated Lennard-Jones interactions between particles. Goals of project:The research aim would be to evaluate the performance and limitations of such a simulation implemented on the Cell processor. Contact:E: Dr Eric McCreath
Wearable Visual Stimulus GeneratorProject Code: CS_06Supervisors:Shaun Cloherty Nick BarnesOutline:Some of the most fundamental questions in visual neuroscience relate to the functional architecture of the visual cortex – that area of the brain charged with processing visual information. In humans, their primate relatives, and some lower mammals, the primary visual cortex is highly organized according to the visual information each cell (neuron) encodes. Although this functional organisation has been thoroughly described, significant questions remain as to how or why the cortex is arranged in this way. It is possible that this structure is innate. It is also possible that the functional organisation may be determined at least in part by the statistics of visual experience during development. Indeed, manipulation of visual input during the critical post-natal period has been demonstrated to influence the tuning properties of individual neurons. Goals of project:This project is aimed at developing a wearable visual stimulus generator to facilitate careful manipulation of visual experience during development in an animal model. This system must be small, light weight and robust – suitable for prolonged deployment in an outdoor environment. Conceptually, this system may include a lowpower embedded microprocessor to drive a pair of small light weight OLED displays, sensors to monitor ambient light levels and activity of the animal, a wireless interface for configuration and monitoring, and appropriate embedded and application software. Discipline:electrical engineering; computer engineering; computer scienceKeywords:embedded systems; wearable computing; wireless sensors; vision scienceContact:E: Shaun ClohertyE: Nick Barnes Laboratory Interface for a Retinal NeurostimulatorProject Code: CS_07Supervisors:Shaun Cloherty Nick BarnesOutline:A neuroprosthesis for the blind - a so-called bionic eye - aims to restore some vision to the profoundly blind by direct electrical stimulation of the surviving retina. However, relatively little is known about how the retina responds to electrical stimulation or how the brain may interpret the signals evoked. To this end, we have undertaken an experimental program to characterise the retinal and cortical response to both visual (light) and electrical stimulation of the retina in an animal model. Goals of project:This project is aimed at developing a laboratory interface for precise electrical stimulation of the retina using microelectrode arrays. Conceptually, this system may consist of a host computer (optically or otherwise isolated from the neurostimulation hardware) presenting to the experimenter an intuitive user interface for configuration of stimulus parameters, an embedded microprocessor to drive a number of software controlled current sources (to be interfaced to the stimulating electrode array), and appropriate embedded and application software. Discipline:electrical engineering; computer engineering; computer scienceKeywords:prosthetic vision; neural stimulation; vision science; embedded systemsContact:E: Shaun ClohertyE: Nick Barnes Caricature and RecognitionProject Code: HCC_1Supervisor:Prof. Tom Gedeon, DCS, CECSOutline:Comparing a specific space to an averaged face allows us to determine how this face differs from the average. If this difference was increased (an inverse of morphing), we can generate a caricature of the face. Caricature faces can be easier to recognise than the original face, since its distinctiveness is boosted. Goals of the project:
Prerequisites:Some image processing / graphics background would be an advantage. References:[1] Making recognisable faces Chatting, D.J.;Automatic Face and Gesture Recognition, 2006. FGR 2006. 7th International Conference on 10-12 April 2006 Page(s):6 pp. DOI 10.1109/FGR.2006.76The full pdf is available if accessed on campus. [2] Face Metamorphosis and Face Caricature: A User's Guide (2004) by Lav Varshney pdf Contact:E: Prof. Tom GedeonView morphingProject Code: HCC_2Supervisor:Prof. Tom Gedeon, DCS, CECSOutline:View morphing between two images of an object taken from two or more different viewpoints produces the illusion of physically moving a virtual camera. This can be used to generate compelling 2D transitions between images. Initially we will use the same face with small rotations and produce intermediate images to produce a realistic movie/QuickTimeVR head. Goals of the project:
Prerequisites:Good mathematical skills, and some experience with image processing. References:[1] View Morphing S. M. Seitz and C. R. Dyer, Proc. SIGGRAPH 96, 1996, 21-30.[2] Automatic target recognition using multiview morphing J Xiao, MA Shah - Proceedings of SPIE, 2004 pdf Contact:E: Prof. Tom GedeonIntegrating EEG and Eye-gaze for image classificationProject Code: HCC_3Supervisor:Prof. Tom Gedeon, DCS, CECSOutline:Human image classification is still relied upon in areas of high significance, such as security image analysis, immigration comparison of passport photos with the live face and so on. This project is to investigate use of eye gaze and EEG (equipment available) to classify images the user is looking at. Goals of the project:
Prerequisites:Experience or an interest in time series analysis or information fusion would be useful, but not necessary. Reference:[1] Construction and Validation of a Neurophysio-technological Framework for Imagery Analysis Andrew Cowell, Kelly Hale, Chris Berka, Sven Fuchs, Angela Baskin, David Jones, Gene Davis, Robin Johnson and Robin Fatch Lecture Notes in Computer Science 4551/2007 pp. 1096-1105Contact:E: Prof. Tom Gedeon
Portable Monitoring of Java Enterprise Applications for Performance ModellingProject Code: e-Gov_01Supervisor:Paul Brebner, Senior Researcher, NICTA e-Government ProjectOutline:The NICTA e-Government project in Canberra is developing a technology for the performance modelling of enterprise software applications implemented as Service Oriented Architectures (SOAs) and n-tier architectures. A common standard and technology stack used to implement such systems is the Java Platform, Enterprise Edition (Java EE, including JSF/JSP, EJB, JDBC, etc). In order to parameterise software performance models, performance data must be captured from selected parts of n-tier applications. The goal of this project is to evaluate approaches for capturing performance data in the form and detail required by a software performance model, in a way which is portable across Java EE products. You will investigate known approaches and products for capturing performance data in Java applications, trial and evaluate selected solutions, and produce a report outlining your discoveries. Prerequisites:Experience, understanding and interest in one or more of: software performance, n-tier software architecture, Java/Java EE programming. Contact:E: Dr Paul BrebnerCharacterisation of Java Enterprise Products for Performance ModellingProject Code: e-Gov_02Supervisor:Paul Brebner, Senior Researcher, NICTA e-Government ProjectOutline:The NICTA e-Government project in Canberra is developing a technology for the performance modelling of enterprise software applications implemented as Service Oriented Architectures (SOAs) and n-tier architectures. A common standard and technology stack used to implement such systems is the Java Platform, Enterprise Edition (Java EE). We are developing a performance modelling approach tailored to work with Model-Driven Development (MDD) of Java EE applications. In order to parameterise software performance models for MDA, the performance characteristics of the target Java EE product must be captured. The goal of this project is to evaluate approaches for capturing performance data in the form and detail required to characterise a selected Java EE product. You will investigate known approaches and products for capturing performance data in Java applications, trial selected solutions, and produce a report outlining your discoveries. Prerequisites:Experience, understanding and interest in one or more of: software performance, n-tier software architecture, Java/Java EE programming. Contact:E: Dr Paul BrebnerDistributed monitoring of Service Oriented Architectures for Performance ModellingProject Code: e-Gov_03Supervisor:Paul Brebner, Senior Researcher, NICTA e-Government ProjectOutline:The NICTA e-Government project in Canberra is developing a technology for the performance modelling of software applications implemented as Service Oriented Architectures (SOAs). Typically SOAs are implemented from a variety of heterogeneous technologies. In order to parameterise software performance models, performance data must be captured from different locations, services and technologies. The goal of this project is to evaluate approaches for capturing performance data in the form and detail required by a software performance model, in a distributed manner, which is portable across SOA implementation technologies. You will investigate known approaches and products for capturing performance data in SOAs, trial and evaluate selected solutions, and produce a report outlining your discoveries Prerequisites:Experience, understanding and interest in one or more of: software performance, software architecture (e.g. SOAs, web services), Java, Web service based applications. Contact:E: Dr Paul BrebnerExploration of the performance characteristics of Orchestration in Service Oriented ArchitecturesProject Code: e-Gov_04Supervisor:Paul Brebner, Senior Researcher, NICTA e-Government ProjectOutline:The NICTA e-Government project in Canberra is developing a technology for the performance modelling of software applications implemented as Service Oriented Architectures (SOAs). SOA applications are typically implemented as workflows of services, which are orchestrated using a variety of patterns and technologies (e.g. centralised workflow engines executing BPEL, Message-based Enterprise Service Buses (ESBs) such as MULE, etc). In order to improve our ability to build accurate performance models of complex SOAs we aim to include explicit models of service orchestration. The goal of this project is to improve our understanding of the performance characteristics of different orchestration approaches, and the means to capture performance data to parameterise the orchestration performance model. You will investigate the performance characteristics of selected orchestration approaches and products, and solutions for capturing performance data, and produce a report outlining your discoveries. Prerequisites:Experience, understanding and interest in one or more of: software performance, software performance testing, software architecture (e.g. SOAs, web services, workflows), Java, Web service based applications, BPEL, ESBs. Contact:E: Dr Paul BrebnerExploration of the performance characteristics of Services and Virtualized ServersProject Code: e-Gov_05Supervisor:Paul Brebner, Senior Researcher, NICTA e-Government ProjectOutline:The NICTA e-Government project in Canberra is developing a technology for the performance modelling of software applications implemented as Service Oriented Architectures (SOAs). SOA services and applications are increasingly hosted on virtual machines (e.g. virtualized servers such as VMWare, Xen). In order to improve our ability to build accurate performance models of complex SOAs we aim to include explicit models of services hosted on virtualized resources. The goal of this project is to improve our understanding of the performance characteristics of virtualized services and applications, and explore the use of virtual servers to isolate services to guarantee Service Level Agreements (SLAs). You will investigate the performance characteristics of selected virtualization products, configurations, SOA applications and SLAs, and produce a report outlining your discoveries. Prerequisites:Experience, understanding and interest in one or more of: software performance, software performance testing, installing and configuring software applications, software architecture (e.g. SOAs, web services, workflows), Java, Web service based applications. Contact:E: Dr Paul BrebnerSOA Portfolio Analysis and ManagementProject Code:e-Gov_06Supervisor:Dr Liam O’Brien, Principal Researcher, NICTA, e-Government ProjectOutline:The use of Service-Oriented Architectures (SOA) [1] approach in building systems is becoming quite popular for many organisations especially government agencies. With the increasing number of services being developed, external services used and applications that use these services there is a need to have portfolio management tools to support better management of all of the various components and artefacts of such systems. The challenge for this project is to examine some of the capabilities of existing portfolio management systems and outline a set of requirements for an SOA Portfolio Analysis and Management tool with visualisation capabilities that makes it easy for IT professionals to get an overview of all of the systems within an organisation and to drill down to detailed information about such systems. Useful capabilities would include for each service what systems it is used in, what are the quality-of-service characteristics and what service level agreements are in place for the service. The challenge within the project is to come up with the set of requirements and if possible a partial prototype of an SOA Portfolio Analysis and Management tool. Prerequisites:Experience, understanding and interest in an organisation’s portfolio of systems and some development experience in Java/Java EE. References:
Contact:E: Dr Liam O'BrienW: Dr Liam O'Brien W: e-Government Project Tools for Data Capture and Manipulation for SOA Project Scope Cost and Effort EstimationProject Code:e-Gov_07Supervisor:Dr Liam O’Brien, Principal Researcher, NICTA, e-Government ProjectOutline:The use of Service-Oriented Architectures (SOA) [1] approach in building systems is becoming quite popular for many organisations especially government agencies. NICTA’s e-Government Project is building a framework for scope, cost and effort estimation. As part of the framework methods and templates are being developed that can be used for scoping and capturing data that is needed in order to support estimation. Data is captured about existing systems, services, stakeholders, development options, integration options, etc. Tools need to be developed that support the capture and manipulation of this data. The challenge for this project is to build some of those tools with functionality that enables capture, visualisation and manipulation of the data. The existing methods and templates can be used as a starting point for the development of support tools and additional methods, templates and tools may be developed. Prerequisites:Experience, understanding and interest in tools development experience in Java/Java EE. Also experience in the use of databases in development would be beneficial. References:
Contact:E: Dr Liam O'BrienW: Dr Liam O'Brien W: e-Government Project
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Please direct all enquiries for this website to:
Page authorised by: Dean, CECS Updated: 27 May, 2009. Page viewed 4743 times. |
|
The Australian National University CRICOS Provider Number 00120C — ABN: 522 34063906 |
