The University of Queensland Homepage
UQ HomepageSearch UQFind all Contacts at UQStudy Options at UQEvents at UQUQ MapsUQ LibraryLink to my.UQ
School of ITEE ITEE Main Website
      





 Thinking Systems - Student Projects Summer 2009

Applications are now closed.

About Thinking Systems

The Thinking Systems Project brings together a cross-disciplinary, collaborative and cross-institutional team to study fundamental issues in how information is transmitted, received, processed and understood in biological and artificial systems. The core of the project is studying how brains understand spatial systems, both physical and conceptual. Insights from neurocognitive systems will be used to develop computational models, autonomous robots and intelligent software agents which in turn will lead to deeper understanding of the relationship between neurocognitive mechanisms and their behaviour in whole systems. The project is jointly funded by the Australian Research Council and the National Health and Medical Research Council.

The following projects are on offer for summer 2009:

We are particularly interested in those students who are considering further study after their undergraduate degree, such as a PhD, and will likely achieve first class honours. This is a good opportunity for you to experience working within a research group. You will be paid a scholarship. Last year the scholarships were from $300-$500 per week. UQ students only.

To apply, or discuss any details, email David Ball with the following:

  • one page c.v.,
  • academic transcript,
  • very short descriptions (1-2 sentences) of how you match the selection criteria,
  • the project(s) you are interested in,
  • your availability during the summer semester (projects will last 8-12 weeks).

The UQ summer break is approximately 12 weeks not including the Christmas period and lasts from the 23rd of November 2009 to the 13th of February 2010.

Applications are now closed.

Project List

SS09-01 - Omni Robot - Docking and Power System (Robotics)

The omni mobile robot platform needs a docking station and power system. The robot must be able to autonomously identify and mate with the dock and recharge its batteries without interrupting the supply to the robot (Similar to how a laptop works.). The robot's power system will distribute clean power from the smart batteries (SMBus) to the actuators (regulated or unregulated), motor drive and sensor boards (5v/3.3V), and motherboard and peripherals (12V/5V).

Selection Criteria

Looking for a mechatronics engineer who has as many of the following practical skills as possible:

  • Mechanical design using CAD packages
  • Embedded electrical design using CAD packages including design, schematics, routing, population, and testing
  • Experience with batteries, chargers, communications buses

SS09-02 - Omni Robot - Omni Vision System (Software)

The omni mobile robot platform needs an omni camera system. The existing system gives a 360 degree view but has problems, particularly from a robot design point of view. One solution is to use multiple (5+) CCD cameras and stitch the images together. You can buy these solutions but they cost up to $20,000. The goal of this project is to develop a cheap, simple multi camera omni vision system.

Selection Criteria

Looking for a software, or mechatronics engineer who has as many of the following practical skills as possible:

  • Experience with writing software to interface to webcams
  • Experience with image processing, in particular, stitching and gain adjustment
  • Experience with embedded software
  • Experience writing drivers

SS09-03 - Omni Robot - Drive and Sensor System (Electrical)

The omni mobile robot platform needs a new electrical motor drive and proximity sensor system. This project will involve selecting a micro controller, designing H-bridges, selecting sensors, and writing embedded software.

Selection Criteria

Looking for an electrical or mechatronics engineer who has as many of the following practical skills as possible:

  • Experience with motor drives
  • Experience with sensors
  • Experience with embedded software

SS09-04 - Omni Robot - Navigation and Behaviours (Software)

The omni mobile robot platform will need to autonomously avoid obstacles and navigate to goals. This project will involve porting our existing navigation system, RatSLAM, to the omni robot. The robot world simulator and the telerobot system will also need to be updated for the new omni robot.

Selection Criteria

Looking for a software or mechatronics engineer who has as many of the following practical skills as possible:

  • Experience with software design
  • Experience with C/C++
  • Experience with robots and navigation

SS09-05 - Robot tracking system (Software)

The lab needs an overhead vision tracking system that can track robots at high speed. Our current system is based on our robot soccer system however the software and interface design is in a hacked state. Taking the ideas from the current system, a new system needs to be designed and written from scratch.

Selection Criteria

Looking for a software engineer who has as many of the following practical skills as possible:

  • Experience with vision systems
  • Experience with C/C++
  • Experience with writing Windows GUI software

SS09-06 - Robot Rat (2 x Mechatronics)

We are in the early stage of designing a robot rat, an advanced micromouse, of the same size as a typical rodent. This project will add the following functionality, dock and recharge, obstacle avoidance, behaviours such as avoidance or tracking of movement and light/dark, cliff avoidance, and navigation.

Selection Criteria

Looking for up to 2 mechatronics engineers who have as many of the following practical skills as possible:

  • Experience with mechatronic systems from sensors to actuators
  • Experience with AI
  • Experience with embedded software

SS09-07 - Rodent Whisking (Mechatronics)

Rodents do a large part of their sensing using their whiskers which can sense distance and texture. The rodent also actively 'whisks' its whiskers along a surface. This project will involve experimenting with various methods of building whiskers and classification of object distance and type.

Selection Criteria

Looking for a mechatronics engineer who has as many of the following practical skills as possible:

  • Experience with sensors
  • Experience with mechanical CAD
  • Experience with actuators
  • Experience with materials engineering

SS09-08 - Wireless Neural Recording (Software)

Biologists require a wireless headset unit for recording electrical signals from a rodent brain. In 2009 we designed a prototype wireless recording system. This project will aim to expand its functionality further including a new GUI interface, integration with overhead tracking, and potentially integration with existing commerical software.

Selection Criteria

Looking for an electrical or software engineer who have as many of the following practical skills as possible:

  • Experience with RF transceivers
  • Experience with embedded software development
  • Experience with Windows GUI software development

SS09-09 - Automatic Behavioural Analysis (Software)

Neuroscientists gather large amounts of data during their experiments. This project will involve automating many of time consuming tasks beginning with the classification of rodent behaviour from an overhead camera offline then potentially in real time. After this we will investigate automation of other parts of the experiments.

Selection Criteria

Looking for a software engineer who has as many of the following practical skills as possible:

  • Experience with visual tracking
  • Experience with MATLAB
  • Experience with C/C++

SS09-10 - Navigation for a High Speed Outdoor Robot (Robotics)

Biologists would like to test theories of navigation on a mobile outdoor robot platform. We already have a small “all terrain” robot car platform. We require the development and testing of simple navigation algorithms using GPS, vision, and range sensor information on this robot platform. The desired end result is a robot car that can use specified sensors to perform simple navigation tasks repeatedly and reliably, in a fully autonomous manner.

Selection Criteria

Looking for a mechatronics engineer who has as many of the following practical skills as possible:

  • Experience coding in Visual C for hardware interfacing between a computer and hardware (i.e. the robot’s camera, sensors such as GPS and cameras)
  • Experience with cameras and image processing
  • Interest in working both in a laboratory and in field tests around UQ

SS09-11 - Boxed In (Modelling)

Neuroscientists have been recording activity from brain regions of rodents and other mammals for many decades. A range of cell types have been identified which seem to be strong candidates for carrying out different navigation tasks. There has been surprisingly rapid consensus that the recently discovered “grid cells” (in the medial entorhinal cortex) subserve path integration in mammals i.e. a kind of vectorial summation process. Sensory and motor noise will corrupt this process leading inevitably to navigation errors. However, in confined experimental arenas, the grid cells display remarkably regular and stable behaviour. We wish to understand the effects of cumulative navigation errors in confined spaces through computational techniques such as Monte Carlo simulations.

Selection Criteria

Looking for a motivated student with strong analytical background with the following attributes (in decreasing order of priority):

  • Experienced with computer simulation techniques, preferably using MATLAB
  • Appreciation of the significance of noise in control systems (including biological nervous systems)
  • Background or strong interest in neurobiology
  • Experience/interest in navigation (animal or robot)

WS09-12 - Lost and Found (Modelling)

There is a belief amongst some neuroscientists that the brain is able to generate optimal solutions to problems. Navigation is both an ancient and ubiquitous function of nervous systems. It is clear that navigation errors occur, but animals have evolved searching strategies to cope. We wish to understand the advantages and disadvantages of various searching algorithms. Under what conditions are they optimal? We wish to consider a range of simple, abstract scenarios which capture the essence of real problems in nature. The goal is to derive or empirically fit functions which quantify the probability of searching success under various conditions.

Selection Criteria

Looking for a motivated student with strong analytical background with the following attributes (in decreasing order of priority):

  • Comfortable with undergraduate level statistics, basic random walk theory
  • Ability for independent in-depth dissection of problems, and development of mathematical proofs
  • Experience/interest in navigation (animal or robot)

SS09-13 - Micro-circuits for oscillation (Modelling)

A major characteristic of brains is that they oscillate. These oscillations span multiple scales and each scale is likely to serve a different type of function. The gamma frequencies (40-100 Hz) are characteristic of sensory regions, with slower beta frequencies used to synchronize between brain regions. We are interested in how these frequencies interact, and transfer information from one time scale and one region to another. We have a candidate micro circuit derived from neural anatomy and need someone with good maths and computing to simulate and test the model.

Selection criteria

Looking for a modelling student who has as many of the following skills as possible:

  • Electrical circuit theory or dynamics
  • MATLAB
  • An ability to pick up new computational designs quickly
  • Some experience or knowledge of neural networks would be useful but is not essential
  • Interest in neuroscience, neural networks, artificial intelligence, robotics

SS09-14 - Human Spatial navigation (Neuroscience)

The project will be in the area of human spatial navigation, with the focus on how humans use and process visual landmarks for navigating successfully through a novel environment. The aim is to gain a better understanding of the fundamental properties of objects as landmarks, the cognitive processes involved in the identification and storage of these landmarks, and its ultimate effect on human navigation. The project will involve behavioural testing of neurologically healthy participants, using a virtual reality paradigm. You will be working with Dr. Oliver Baumann and Professor Jason Mattingley at the Queensland Brain Institute in Prof. Jason Mattingley's Cognitive Neuroscience Laboratory. If you would like further information, please feel free to email (o.baumann@uq.edu.au).

Selection criteria

Looking for a neuroscience student who has as many of the following skills as possible:

  • Behavioural testing
  • Excel
  • SPSS (Statistical Package for the Social Sciences)

SS09-15 - Computational Capacity (Software, Algorithms)

The goal of this project is to increase the computational capacity of the group -- speed up our existing algorithms and simulations --. A wide variety of technologies may be utalised including:

  • Multi-cores (8+)
  • Serverfarms
  • CUDA (Nvidia's GPU tookit)
  • Hardware implementations

This is a broad project with many potential avenues for investigation. The first stage would involve a preliminary investigation of different approaches for different projects.

Selection Criteria

Looking for an engineer or computer scientist who has as many of the following skills as possible:

  • multi-core, multi-pc computing
  • MATLAB
  • C/C++
  • makefiles
  • CUDA
  • Computational Modelling

SS09-16 - Attention Oscillation Model (Computational Modelling)

Biologists require a model of oscillatory activity to look at neural mechanisms of visual attention.

The first goal is to implement a computational model of the insect brain network based on neural circuitry to reproduce the oscillations observed in the brain. Next, different portions of the network would be manipulated to determine what areas are required to produce these oscillations. The modelled activity of individual modelled neurons would be compared with the network behaviour to determine whether the single unit activity corresponds with the population. The model could be tested to determine whether the network can be wired to ‘attend’ to conflicting ‘stimuli,’ based on the connections within the network.

Selection Criteria

The student should have the following experience:

  • MATLAB/C/C++
  • Computational Modelling, Neural Networks

SS09-17 - Modeling of scent processing in the honeybee - how are complex scents recognised? (Computational Modelling)

Smell (olfaction) is the oldest sense, possessed even by simple creatures with no eyes or ears. Consequently, olfactory receptors connect to very deep parts of the brain, and the functioning of the olfactory sense can reveal fundamental mechanisms of brain operation. If we can understand how the sense of smell works, we can apply what we have learned to help understand many other parts of the brain.

Recent work by scientists at the University of Queensland has revealed new principles for the inner workings of the sense of smell. We would like to build an artificial neural network model to demonstrate these principles. The ideal candidate for this project will have an interest in neuroscience and good programming ability.

Selection Criteria

The student should have the following experience:

  • MATLAB/Python
  • Computational Modelling, Neural Networks

SS09-18 - Adding objects to spatial language games (AI)

The Lingodroids project investigates fundamental questions about the evolution of language, using agents in grid worlds, as well as simulated worlds and the real world with Pioneer robots. In a typical language game, the mobile robots explore and map their world. They then play language games with each other to name different areas and spatial relations, learning parts of speech corresponding to nouns and prepositions. The complexity of the languages they develop depends on the richness of their internal representations and the complexity of the language games.

This project aims to extend the concepts formed by the agents to descriptions of views or objects located in the world. This will allow the agents to talk about what they can see at various locations in the world, including locations which have only been visited by one of the agents. Preliminary games have been developed in the grid world. The aim will be to run simulations that demonstrate the ability of the agents to talk about features of their world.

Selection Criteria

The student should have the following experience:

  • Good programming skills
  • Willing to learn about language games, linguistics and object representation

SS09-19 - Extending physical spatial language games to colour space (AI)

The Lingodroids project investigates fundamental questions about the evolution of language, using agents in grid worlds, as well as simulated worlds and the real world with Pioneer robots. In a typical language game, the mobile robots explore and map their world. They then play language games with each other to name different areas and spatial relations, learning parts of speech corresponding to nouns and prepositions. The complexity of the languages they develop depends on the richness of their internal representations and the complexity of the language games.

This project aims to apply the language game structures developed to name locations and spatial relations to a different domain, specifically the domain of colour. The project will involve the development of representations for colour terms that the agents will be able to use, and the development of games for the agents to use to form concepts in colour space. The aim will be to run simulations that demonstrate the ability of the agents to form concepts for colours and relationships between colours.

Selection Criteria

The student should have the following experience:

  • Good programming skills
  • Willing to learn about language games, linguistics and colour representation
© 2002-2009 The University of Queensland, Brisbane, Australia
ABN 63 942 912 684
CRICOS Provider No: 00025B
Authorised by: Head of School
Maintained by: webmasters@itee.uq.edu.au
Last Updated: 12 October, 2009, 5:59pm