Information Capability PhD Project List

Interested in research in information capability at the University of Adelaide? This outstanding research training program is fully funded through a University or industry scholarship.

This will give you the opportunity to develop deep expertise in a subject area that is of critical importance to future enterprises, both public and private. Much of our research in information capability is being conducted in partnership with industry, in the defence, public and commercial sectors. 

To view the current projects available, please check the list below. 

Once you’ve found a project you’re interested in, please email the relevant contact for that project and provide a copy of your CV and academic transcripts. Suitable candidates will be invited to apply by the supervisor, and additional information regarding how to submit their formal application will then be provided by the Graduate Centre.

For additional information on eligibility and the scholarship please click here. Some projects may also be restricted to domestic applicants or Australian citizens only. Please consult each project description below or the academic contact for more information.

Applications close 9 October 2020.

    Expand
  • Social influence and policy

    How can influence activities be integrated into Australian strategic analysis?

    Project ID: ARTS-01
    Field of research: Politics and International Relations

    With influence activities in the digital domain increasing, Australian strategic analysis needs to account for the full range of ways in which influence activities may modify Australia’s behaviour, or how Australian influence activities can influence other states’ behaviour. There is an urgent need to develop a more nuanced understanding of what influence means, how influence activities in the information environment interact with traditional levers of influence, and how influence may be operationalised in pursuit of Australia’s strategic interests.

    Supervisor(s): Prof Joanne Wallis
    Contact: Prof Joanne Wallis; joanne.wallis@adelaide.edu.au


    What are the implications for Australia of influence activities in its immediate region?

    Project ID: ARTS-02
    Field of research: Politics and International Relations

    While there is much talk in Australian strategic and foreign policy debates about influence activities in Australia’s immediate region, there has been minimal analysis of what these activities achieve in practical terms. That is, do these activities actually influence the decisions of politicians, other public decision-makers, or populations more broadly? How can we assess this? How do influence activities in the information environment interact with more traditional levers of influence, such as aid, military assistance and investment? What are the implications for Australia?

    Supervisor(s): Prof Joanne Wallis
    Contact: Prof Joanne Wallis; joanne.wallis@adelaide.edu.au


    Modelling and predicting the spread of (mis)information online

    Project ID: ECMS-10
    Field of research: Applied mathematics

    How do things “go viral”? Can this be predicted? This project will develop new models for the “complex contagion” of ideas, and examine how they might be used to predict the spread of information, misinformation, and emotions, through online social networks. Using data from some of the largest social media platforms, you will study how conspiracy theories can spread and be amplified.

    Supervisor(s): Dr Lewis Mitchell
    Contact: Dr Lewis Mitchell; lewis.mitchell@adelaide.edu.au

  • Space science and technology

    Physics; laser photonics

    Project ID: SCI-01
    Field of research: Physics; laser photonics
    Notes: Available to Australian citizens only

    Supervisor(s): Prof Peter Veitch
    Contact: Professor Peter Veitch; peter.veitch@adelaide.edu.au


    Understanding the Law Applicable to Information Warfare

    Project ID: PROF-01
    Field of research: Law and Politics

    Supervisor(s): Prof Melissa de Zwart, Assoc Prof Tim Legrand
    Contact: Prof Melissa de Zwart; melissa.dezwart@adelaide.edu.au


    Security and Supply-Chain Control of Networked Circular-Looped Systems - for Manufacturing in Space and on post-COVID Earth

    Project ID: ECMS-11
    Field of research: Space, Communication and Network Engineering
    Notes: Partnership is sought with: DST (Network security), MineARC-Biora and AgGrow Energy Resources (greenhouse tech), AstroHub Pte (space-STEM education)

    Post-COVID manufacturing moves towards small-scale, local, resilient and self-sufficient manufacturing - circular-looped systems. The supply chains of this new ‘fractal economy’ and its future factories need to be trimmed towards the same attributes, and need to be protected by robust, cyber-defence systems. The multiple communication and control of the process mass and energy flows and resource transportation, at sites all over Australia, demands for advanced network engineering and dimensioning. We will develop network protocols and dimensioning solutions to meet the performance and security needs of these remote manufacturing systems.  

    As a challenging example for such complex system, the information capability of the recently developed ‘space greenhouse’ and its integrated coupling to renewable energy sources and resource extraction will be considered – this own preparatory work can provide an experimental mock-up with several electronic functionalities (including microgravity operation), sensors for plant and nutrition testing, and a TRC-designed supply-chain modelling platform for moon-resource mapping. Circularity indicators (CTIs) will assess the sustainability, e.g. through the ‘zero-waste economy’ by the material circularity indicator (MCI) and the avoidance of critical resource supply monopoly by the Herfindahl-Hirschman Index (HHI).

    Supervisor(s): Dr. Hung Nguyen, Co-supervisor Prof. Volker Hessel
    Contact: Dr. Hung Nguyen; hung.nguyen@adelaide.edu.au 

  • AI and cybersecurity

    Wireless Cyber-Physical Information Networks

    Project ID: ECMS-02
    Field of research: Computer Science
    Notes: This is a joint project, co-supervised with DST

    Whilst much attention is given to high-level aspect of cyber attacks and defence for information networks, there has been recent interest in the vulnerabilities of lower level physical systems such as software defined radio platforms. With the increasing deployment of tactical cognitive wireless information networks, it is on considerable importance to understand vulnerabilities of such networks from the point of view of protection and exploitation. This project will examine cross-layer design issues for such systems from a cyber perspective.

    Supervisor(s): Prof Lang White
    Contact: Professor Ian Reid; ian.reid@adelaide.edu.au

    Additional criteria: This project is only available to Australian citizens. 


    Intelligent Technologies for Smart Cryptography

    Project ID: ECMS-12
    Field of research: Computer Science

    Smart devices are the key enabler of the smart society we live in.
    They are the building blocks which construct smart homes, smart cars, smart grids, smart cities, and even smart governments. Research has uncovered severe flaws in the cryptographic substrate in smart devices, and several aspects of smart devices make them extremely difficult to protect: (1) they are often resource-constrained, (2) smart devices are often distributed to untrusted end users, thus allowing physical attacks, and (3) in many cases time-to-market and development cost considerations drive smart device development.

    • Among other, possible research directions include (1) the efficient generation of cryptographic code from high-level descriptions, (2) accurate modelling of information leakage of concrete devices, and (3) automatic elimination of information leakage.
    • Interdisciplinary projects (ex CompSci) are possible with anyone who has an interest in keeping information private, ranging from chip manufacturers to providers of large, distributed systems of smart devices.

    Supervisor(s): Markus Wagner, Jason Xue, Chitchanok Chuengsatiansup and Yuval Yarom
    Contact: Dr. Yuval Yarom yval@cs.adelaide.edu.au


    AI-Enabled Platform for Validation and Fusions of Security Threat Data

    Project ID: ECMS-13
    Field of research: 
    Notes: 
    Partners: Splunk, SA Health and Trusted Security Services (TSS)

    The aim of this project is to design, implement and evaluate an Artificial Intelligence (AI)-Enabled framework for validating security threat data coming from different sources and an integrated tool suite. The envisioned platform will provide advanced AI (ML/DL/NLP/Sematic) algorithms and their implementations for automatically categorizing and integrating security threat data for supporting superior decision making by security staff working in a Security Operations Centre (SOC). The security tool suite hosted by the platform will provide their services through suitable Applications Programming Interfaces (APIs) for seamlessly integrating and operating with the existing and new Security Information and Events Management (SIEM) and Endpoint Detection and Response (EDR) platforms for effective and efficient operations of a SOC.

    Supervisor(s): Prof Ali Babar, and TBA
    Contact: Professor Ian Reid; ian.reid@adelaide.edu.au


    User Behavioural Analytics for Healthcare Security

    Project ID: ECMS-14
    Field of research: 
    Notes: 
    Partners: SA Health and Splunk

    The stakeholders of e-health care systems often usually lack security awareness and security expertise. In such situations, the actions of a healthcare stakeholder (e.g., patient or hospital operator) can lead to massive data breaches, which can compromise the provision of provision of health services securely and reliably. This project aims to collect the large volume of operations logs from various e-health systems and use the logs for profiling user behaviours. The project will research on developing and evaluating advanced big data analytical and visualisation approaches using Machine Learning and Deep Learning approaches for incremental update and detecting deviations from the usual behaviours.  The key outcome of the project will be a set of algorithms and a tool suite for collecting and analysing user logs for building and leveraging user profiles for enhanced security of healthcare systems.

    Supervisor(s): Prof Ali Babar
    Contact: Professor Ian Reid; ian.reid@adelaide.edu.au


    Engineering Secure ML-Enabled cybersecurity systems

    Project ID: ECMS-15
    Field of research: 
    Notes: 
    Partners: DST, Data61

    Machine Learning (ML) enabled cybersecurity systems such as Intrusion Detection Systems (IDS), phishing detectors and spam filters have gained the attention of both research and industrial community. Unlike the traditional software systems, little known about how to engineer ML-enabled cybersecurity systems. This project will develop a model-driven framework for enabling the development of secure ML-enabled cybersecurity systems. Specific research contributions can be made to developing approaches for logging and auditing ML development lifecycle to enable security traceability in these systems, penetration testing of ML based security systems, impact of poisoning attacks on ML-based cybersecurity systems, or developing robust ML-based cybersecurity to enable their practical applicability.

    Supervisor(s): Prof Ali Babar
    Contact: Professor Ian Reid; ian.reid@adelaide.edu.au


    Ensuring Privacy Preserving AI Technologies

    Project ID: ECMS-16
    Field of research: 

    The increasing use of AI technology requires viable ways for stakeholders to share datasets and models in a privacy-preserving manner, respecting the privacy of individuals as well as intellectual property rights. The adoption of private AI mechanisms can increase the accuracy of existing solutions and foster collaboration across industry and academia. This project is aimed at building and evaluating innovative methods and tools for implementation and evaluation of privacy-preserving AI technologies, protecting datasets and machine learning models in multi-owner, outsourced and collaborative ecosystems.

    Supervisor(s): Prof Ali Babar
    Contact: Professor Ian Reid; ian.reid@adelaide.edu.au


    Context Aware Network Vulnerability Analysis with Metagraphs

    Project ID: ECMS-17
    Field of research: Communication and Network Engineering
    Note: This is a joint project, co-supervised with DST

    Multi-step network intrusion is the dominant methods used by sophisticated attackers to infiltrate networks. Carefully crafted multi-step attacks can sidestep defences and turn trivial security flaws into major holes. Currently attack graphs are used to analyse potential multi-step attacks. These works first build a graph of potential attacks and then apply various techniques to iterate through all the potential attack paths.  The volume cyber-security vulnerabilities is constantly growing. Protecting against all combinations of these vulnerabilities is impractically costly. For these reasons, most current attack graph solutions do not scale well and are too slow for real-time cyber defence. We will explore how to use abstract Metagraph algebra to integrate specific network context to reduce the complexity of attack graphs. Using these real-time Metagraph based attack-graphs, we can rapidly analyse the current situation and possible actions to guide AI/game theoretic countermeasures against proactive adversaries. 

    Supervisor(s): Dr Hung Nguyen (TRC, UoA), Prof. Matthew Roughan (Maths, UoA), Dr Peyam Pourbeik (CEWD, DST)
    Contact: Professor Ian Reid; ian.reid@adelaide.edu.au


    Real-Time Network Segmentation and Segregation using Graph Convolutional Neural Networks

    Project ID: ECMS-18
    Field of research: Communication and Network Engineering
    Note: This is a joint project, co-supervised with DST

    Network segmentation involves partitioning a network into smaller networks, while network segregation involves enforcing security rules to restrict access to hosts and services only when required and authorized. Through network segmentation and segregation, the access to sensitive information, hosts and services can be restricted without affecting the organization and operation of the business. Network segmentation and segregation is one of the most effective techniques to prevent attacks and limit their impact. Today network segmentation and segregation are mainly done manually where security teams follow some guidelines that are mostly generic and require a high level of expertise. Building on our recent work in graph partitioning algorithms and auto-configuration, we will investigate automated methods for real-time network segmentation and segregation. Most current AI solutions  often indiscriminately block all network traffic on all ports and connections, rather than considering the role of each machine and the motive behind the compromise attempts as a human defender might. These solutions thus unnecessarily disable communications that are key to business operations. The main idea of the project is to develop segmentation and segregation methods that ensure operation continuity by creating a system of prevention and recovery from potential threats to the network infrastructure.

    Supervisor(s): Dr Hung Nguyen (TRC, UoA), Prof. Matthew Roughan (Maths, UoA), Dr Peyam Pourbeik (CEWD, DST)
    Contact: Professor Ian Reid; ian.reid@adelaide.edu.au


    Provable Network Security with Mechanism Design

    Project ID: ECMS-19
    Field of research: Communication and Network Engineering

    Security games provide an analytical framework for modelling the interaction between the attacker, who aims to access a particular network resource, and the defensive agent who tries to stop the attacker. Mechanism design is a field in game theory that takes an objectives-first approach to designing mechanisms toward desired objectives. In this project, we will explore how to apply mechanism design to build network and security policies that guarantee that the defender always succeeds in securing the network and critical data. Of particular interest is the Gibbard–Satterthwaite impossibility theorem - that for a general class of games, only "dictatorial" social choice functions can be implemented. That is, it’s only possible to design a game where one agent always receives his most-favoured goods allocation. This theorem is among the most remarkable negative results in economics. We will investigate if we can use this theorem to the network defender’s advantage to design network and security policies where he is the “dictator”.

    Supervisor(s): Dr Hung Nguyen
    Contact: Professor Ian Reid; ian.reid@adelaide.edu.au


    Hannan Consistent Countermeasures Against Non-Stationary Adversarial Attacks on Autonomous Cyber Operations

    Project ID: ECMS-20
    Field of research: Communication and Network Engineering
    Notes: This is a joint project, co-supervised with DSTG

    Autonomous Cyber Operations (ACO) is a disruptive new technology that will have profound impact on Defence and industry. Game theory has been proposed as the most promising framework to address the problems of proactive adversaries in ACO systems. Most current game theoretic solutions, however, assume that the attackers employ a rational best-response strategy to make their decisions. The performance of the games when attackers deliberately choose an arbitrary, non-optimal, non-stationary strategy is not well understood. We will develop novel game theoretic methods that guarantee to deliver secure cyber systems regardless of the strategy of the adversary in this research. In order to obtain this desired outcome, we will develop a novel game solution concept for ACO using Hannan consistent (or external regret minimising) strategies. The goal of external regret minimising algorithms is to ensure that the time average regret approaches zero, regardless of the opponents’ strategies.

    Supervisor(s): Dr Hung Nguyen
    Contact: Professor Ian Reid; ian.reid@adelaide.edu.au

  • Autonomy and robotics

    Bio-inspired soft robot for lunar and planetary mission

    Project ID: ECMS-21
    Field of research: 

    Soft robots, emerging type of robots, offer higher dexterity and flexibility than the typical rigid robots that usually come with joints. These two important capabilities are critical in adapting with unstructured and cluttered environments, so that better manipulation and locomotion can be achieved. In lunar and/or Mars, the low gravitational forces will affect the optimal rate of motion, alter the centre of mass movements and gait activity patterns. The project aims to design multi-tasking soft robotic platform from soft that is suitable for such condition, thus can be utilized for lunar and planetary mission. There are a number of aspects which will be considered in this project; 1) the materials for the soft robot that can stand the harsh environment conditions with extreme temperature differences and ionic radiation, 2) the platform design, which  will be based on what nature has offered, such as the octopus with its eight tentacles, and/or locust with its jumping and flapping capabilities, hence it should be able to perform highly flexible locomotion that suits this uncertain and extreme condition, and 3) the control system for such robotic platform.     

    Supervisor(s): A/Prof. Rini Akmeliawati, Dr. David Harvey and A/Prof. Ling Yin
    Contact: Professor Anthony Zander; anthony.zander@adelaide.edu.au


    Autonomous Flight Control Systems Insect-mimicking Flapping Wings Micro Aerial Vehicles

    Project ID: ECMS-22
    Field of research: 

    The project aims to design and create an automatic stabilization and control system of an insect-inspired flapping wing micro aerial vehicle (FWMAV). This type of micro aerial vehicles has complex, periodic, time-varying and inherently unstable dynamics. To date, their flight dynamics present very challenging control problems. In this project, energy-based control approaches will be exploited and created to provide automatic stabilization and control for beetle-mimicking FWMAV. 

    Supervisor(s): A/Prof. Rini Akmeliawati, A/Prof. Steven Grainger, Prof. Iven Mareels (University of Melbourne)
    Contact: Professor Anthony Zander; anthony.zander@adelaide.edu.au


    Distributed Control of Modular Aerial Vehicles

    Project ID: ECMS-23
    Field of research: 

    The study aims to develop a distributed control strategy for multi-rotor platform of modular UAVs. The multi-rotor platform consists of several modular unmanned aerial vehicles. Every single vehicle, in addition to coordinating and communicating with its peers, can fly either independently or within a group to form a single entity after on-air docking with the other vehicles. Therefore, the distributed control approach will be used to manage the configuration of the vehicles covering the different aspects of multi-rotor vehicles including on-air docking and flying to achieve a better performance and to better enhance the surveillance and rescue operations tasks.

    Supervisor(s): A/Prof. Rini Akmeliawati, A/Prof. Steven Grainger, Dr. Nataliia Sergiienko
    Contact: Professor Anthony Zander; anthony.zander@adelaide.edu.au


    Energy-based Control and Navigation Systems of Autonomous Underwater Vehicle

    Project ID: ECMS-24
    Field of research: 

    In this project, an integrated energy-based and visual servo control system is proposed for an autonomous underwater vehicle (AUV) to provide automatic navigation and control as well as 3D image reconstruction of sea environment. The proposed energy-based and visual-servo control provides integrated stabilization, motion control and navigation. Energy-based control is proposed as it allows for the use of the energy function to model and design the controller, which can provide guaranteed robust stability and performance. Visual-servo control relies on the image data obtained in real-time from which the position of the vehicle can be estimated, and obstacle avoidance and trajectory tracking can be achieved. The image data used for the controller together with the data from other sensors, such as sonar and laser vision sensors, can be further utilized for mapping and 3D reconstruction of the undersea environment, which can be also used to define the path for the navigation system.

    Supervisor(s): A/Prof. Rini Akmeliawati, Dr. Eric Fusil
    Contact: Professor Anthony Zander; anthony.zander@adelaide.edu.au


    Bio-inspired Autonomous Robots

    Project ID: ECMS-38
    Field of research: Research as Neuroscience and Robotics 

    Animals have evolved relatively simple and efficient solutions to tasks that challenge the most sophisticated robotic vision systems. For example, flying insects can detect and pursue moving features (prey) in cluttered environments, predicting their future location and selecting one target amidst a swarm. Using electrodes 1500 times thinner than the width of a human hair, we record from neurons in the insect brain that underlie this ability to encode the surrounding visual environment. Using insights from both insects and other animal model systems, we develop bio-inspired, computational models and translate these onto our autonomous ground and aerial vehicles. These autonomous systems navigate the world, emulating biological behaviours. This project therefore crosses the disciplines of neuroscience (Adelaide Medical School) and robotics (School of Mechanical Engineering). As this research involves computational modelling and hardware development, it is suited to those with mathematical or engineering backgrounds. 

    Supervisor(s): A/Prof Steven Wiederman, Prof Ben Cazzolato and A/Prof Steven Grainger 
    Contact: Professor Anthony Zander; anthony.zander@adelaide.edu.au


    Autonomy and Robotics (UAV and/or UGV communication, collaboration, formation, detection, tracking, consensus, control)

    Project ID: ECMS-05
    Field of research: Electronic Engineering

    Multi-Agent Systems (MASs) have the characteristics of cooperation and decentralization. As the agents often work under complex circumstances, limitations of the hardware that include limited passive sensing and active communication capabilities are likely to be present. Agents need to cooperate in a distributed manner. Formation control of multi-agent systems has been widely used for cooperative tasks in such applications as terrain exploration, mobile networks and traffic control. However, the communication-induced problems and the high failure risk of increasing equipment has created a number of challenges for the security of MASs. The objective of this project is to design a robust formation control strategy for a multi-vehicle system against communication/physical failures (e.g., network attacks, link failures, packet dropouts, sensor/actuator faults). The problems to be investigated in this project include methodologies for formation control and self-assembly for multi-vehicle systems; and agent-based systems for target searching and delivering.

    Supervisor(s): Prof Peng Shi
    Contact: Professor Anthony Zander; anthony.zander@adelaide.edu.au


    Cyber-physical systems (control, resilience, robustness, reliability, securities against attacks, such as disruption, deception and disclose)

    Project ID: ECMS-06
    Field of research: Electronic Engineering

    With the deep integration of control, communication, computation, cloud and cognition, cyber-physical systems (CPSs) act as the backbone of Industry 4.0. A CPS is a mechanism that is controlled or monitored by computer-based algorithms, with tight interconnection with the Internet and its users. CPSs has wide applications, such as intelligent transportation, medical health-care device, and power grid. Due to the intrinsic vulnerability of the cyber layer, it is non-negligible that CPSs may suffer from different adversarial attacks with the purpose of degrading system performances. Thus, it is important and critical to develop methodologies to protect the security and reliability of CPSs. In this project, the candidates will focus on investigating defence-based control system design to realize the attack resilience. Some relevant research topics include secure estimation, attack detection and identification, optimal attack-resilience control and with the potential applications on power systems.

    Supervisor(s): Prof Peng Shi
    Contact: Professor Anthony Zander; anthony.zander@adelaide.edu.au


    Network systems (control, communication, delays, data drop-off, uncertainties, attacks)

    Project ID: ECMS-07
    Field of research: Electronic Engineering

    With the development of computer, communication and control technologies networked control systems (NCSs) have been widely used in critical infrastructures such as power systems, chemical industries, manufacturing, transportation management, and teleoperation systems. Signals of control systems are transmitted via networks, which increases flexibility, interoperability and resource sharing of a control system. However, the use of communication networks makes NCSs not only subject to network-induced communication constraints such as time-delays, packet disorders and dropouts but also susceptible to malicious cyber-attacks, both of which may lead to poor system performance and even system instability. You will develop novel and reliable modelling, analysis and design methodologies for NCSs to maintain desired system performances.

    Supervisor(s): Prof Peng Shi
    Contact: Professor Anthony Zander; anthony.zander@adelaide.edu.au


    Intelligent systems (pattern recognition, imaging processing, data processing)

    Project ID: ECMS-08
    Field of research: Electronic Engineering

    Intelligent systems are technologically advanced machines that perceive and respond to the world around them. Intelligent systems can take many forms, from automated vacuums to personalized reading and shopping suggestions, and from driver-less vehicles to information processing. In this project, you will study problems of pattern recognition. Real-time object recognition is a technology of identifying object with fast inference and maintaining a high level of accuracy. You will develop novel and reliable pattern recognition techniques (by machine learning, neural networks, and artificial intelligence), which will be tested by simulations and real experiments.  

    Supervisor(s): Prof Peng Shi
    Contact: Professor Anthony Zander; anthony.zander@adelaide.edu.au


    A framework for digital engineering of future systems: from strategy to systems architecture

    Project ID: ECMS-25
    Field of research: Digital Engineering

    Systems design is moving towards a digital engineering implementation, where traditional document-based and disconnected approaches are replaced with a digital, connected approach. Digital engineering sees key design artefacts linked to digital data and models and an information-centric and evidenced-based approach to design. This project will explore how digital engineering can evolve to be support the early-stage design of non-deterministic and highly connected systems with unknowable emergent behaviours. This is an industry-linked project in collaboration with Shoal. The successful applicant will work with Shoal to explore future design considerations, design the framework and refine and test it on a pilot project developing the high-level design of an autonomous system for an identified strategic purpose.

    Supervisor(s): Professor Stephen Cook, Dr David Harvey and Mr Kevin Robinson
    Contact: Professor Anthony Zander; anthony.zander@adelaide.edu.au

  • Sensors and devices

    Vehicle and built features classification of Synthetic Aperture Radar images using machine learning

    Project ID: ECMS-26
    Field of research: Electrical & Electronic Engineering

    Synthetic aperture radar (SAR) is a form of radar that is used to create two-dimensional images or three-dimensional reconstructions of objects, such as landscapes. It is widely deployed for military and civilian applications, such as environmental and large-scale soil moisture monitoring. Current generation of space and airborne SARs are capable of capturing vast quantities of data at high resolutions. An effective classification system has the potential to greatly expand the usage of datasets and add great value to the SAR datasets. Military examples include wide area surveillance where multiple targets may overwhelm human operators. In civilian applications, reliable classification will enable (near) real-time monitoring of infrastructure. With increasingly large datasets available, it has become feasible to apply data-intensive techniques, such as machine learning, to the analysis of image data. 

    This project is concerned with the detection and classification of objects of interest in high resolution radar imagery. Objects of interest are primarily moving or stationary vehicles, but it can also incorporate roads, houses and other forms of man-made structures. The research will leverage on recent advances in machine learning, such as 'deep learning' structures, to improve the systems classification performance. The Adelaide Radar Research Centre has a long history in remote sensing applications using SAR and has a large existing dataset to support the proposed research.

    Supervisor(s): Associate Professor Brian Ng 
    Contact: Associate Professor Brian Ng brian.ng@adelaide.edu.au

    Additional criteria: This project is only available to domestic applicants. 


    Dictionary learning techniques on radar imagery

    Project ID: ECMS-27
    Field of research: Electrical & Electronic Engineering

    Recent advances in radar systems have led to large volumes of data. Finding the relevant piece information from the sea of data is a `needle in the haystack’ exercise but one that is increasingly important for many applications such as surveillance and tracking. It is impossible for human operators to perform these tasks with sufficient speed to keep up, hence effective automatic methods are sorely needed. Despite impressive advances in other fields, artificial intelligence methods have not been widely applied to radars. One reason may be the unique characteristics of radar imaging methods that distinguish their signals from other sensing modes. On the other hand, dictionary learning, often coupled with sparse representation approaches, have shown promise in a variety of radar applications, from interference suppression to target recognition. 

    This project aims to explore novel dictionary learning algorithms on synthetic and inverse synthetic aperture radar images. It is envisioned that advances made will lead the state-of-the-art towards the `holy grail’ of automatic target recognition. The Adelaide Radar Research Centre works closely with DST Group and will be able to access high quality real-world datasets for this research.

    Supervisor(s): Associate Professor Brian Ng 
    Contact: Associate Professor Brian Ng brian.ng@adelaide.edu.au

    Additional criteria: This project is only available to domestic applicants. 


    Information Capability: Antennas for Electromagnetically and Physically Challenging Environments

    Project ID: ECMS-39
    Field of research: Electronic Engineering

    Modern radio communications are placing increasing demands on the performance of antennas as critical components of wireless systems. State-of-the-art implementation increasingly requires flexible operation offering a variety of functionalities over wide frequency ranges. While such diversity aims to increase the reliability of communications within often contested electromagnetic and challenging physical environments, it equally offers the engineer the potential of creative design according to both a fundamental and an applied understanding of electromagnetic theory. This project will investigate current advances in wireless design to develop compact antennas capable of operating in several wideband radiation modes that coexist with limited interaction in a constrained volume. A primary objective of this research is to contribute to the state-of-the-art through novel design techniques that offer the potential to innovate in highly relevant areas such as beam switching, beam scanning, smart reconfiguration and electrically small design.

    Supervisor(s): Prof. Christophe Fumeaux, Dr. Nicholas Lawrence (DST Group)
    Contact: Prof. Christophe Fumeaux, christophe.fumeaux@adelaide.edu.au

    Additional criteria: This project is only available to Australian citizens. 

  • Human autonomy teaming

    Dialogue analysis and management using game theoretic analyses of information exchange

    Project ID: ECMS-01
    Field of research: Engineering and Psychology

    The objective of this project is to manage the dynamics of dialogue exchange between humans and AI agents. The project will focus on modelling and tuning the iterative content of exchanges between agents, so as to reach a satisfactory state of shared knowledge whilst allowing each agent to execute its respective tasks (either on its own or in conjunction with other agents). The project involves an exciting combination of game theory, information theory, linguistics and behavioural analyses.

    Supervisor(s): Prof Lang White, A/Prof Anna Ma-Wyatt, Dr Conrad Perry
    Contact: A/Prof Anna Ma-Wyatt; ann.mawayatt@adelaide.edu.au


     

    Integrating eye tracking with SLAM to understand attention in a real environment

    Project ID: HMS-01
    Field of research: Psychology and Computer Science

    People use visual information as they interact with the world, and there is considerable lab based evidence that the eye movements people make play a key role in deploying attention. However, there is currently limited information about how people deploy attention in a real environment and while there are advanced computer vision approaches for identifying objects in real time, semantic segmentation is still limited. In this project, we will integrate real world eye tracking with SLAM to investigate how people pay attention in a real environment. This approach will advance our understanding of what parts of a scene people pay attention to as they conduct tasks and navigate environments, and also understand how deployment of attention over time helps people build up a representation of meaning in the scene. These outcomes will be relevant for neuroscience and computer vision, and relevant for the development of advanced sensing systems and information systems.

    Supervisor(s): A/Prof Anna Ma-Wyatt, Prof Ian Reid
    Contact: A/Prof Anna Ma-Wyatt; ann.mawayatt@adelaide.edu.au


    Embedded AI platform to push intelligence to the edge or front-end

    Project ID: ECMS-03
    Field of research: Computer Science

    Many field works (such as farming and fishing) cannot carry a super-computer to compute and inference in real-time. And using cloud computing is not an option due to the low broadband connectivity in regional areas. Hence, we need smart edge computers in a sensor with AI chips, which enjoy computational efficiency for lower power. This also opens the door to turn an existing online or computation-cost system into a handheld device for regions.

    Supervisor(s): Prof Javen Shi and Dr Ehsan Abbasjenad
    Contact: Professor Ian Reid; ian.reid@adelaide.edu.au 


    Learning and reasoning framework for sparse, digital unfriendly, and multi-modal data

    Project ID: ECMS-04
    Field of research: Computer Science

    Much data and knowledge in the regions are not digitalised (handwritten notes and receipts) or digitalised in an unfriendly way (a report being scanned and stored as images instead of texts, voice memos). Moreover, a single application may involve data and knowledge in multiple modalities (e.g. texts, images, videos, spectra, sound, database, and knowledge bases). We find this is a practical problem for every data scientist on every project. Hence an Open API could benefit multiple governments and industry partners. Another emerging issue is some critical data collected are very sparse due to cost, time and other restrictions, for example, in mining. Thus, we need a transferable meta-learning method to extract and learn the metadata so that we can start to reason without the need for re-training the reasoning module.

    Supervisor(s): Prof Javen Shi and Dr Ehsan Abbasjenad
    Contact: Professor Ian Reid; ian.reid@adelaide.edu.au 


    Learning to continually learn

    Project ID: ECMS-37
    Field of research: Computer Science

    A fundamental aspect of intelligence is learning continuously from new observations. Current machine learning algorithms, however, are far from utilising new data for updating the model which pose as a significant limitation. In particular, in interactive applications such as robotics or dialogue systems where the complete state of the world is generally unknown beforehand, it is crucial to be able to learn on the fly and update the existing model. In this project, we focus on continual learning for agents to be able to interact with the world and learn from new observations.

    Supervisor(s)Dr Ehsan Abbasjenad and Prof Javen Shi
    Contact: Professor Ian Reid; ian.reid@adelaide.edu.au 

     

  • Artificial intelligence and machine learning

    Novel postgraduate research into computer vision, machine learning and deep learning

    Project ID: AIML-01
    Field of research: Machine Learning

    The Australian Institute for Machine Learning (AIML) is ranked in the top three of global research organisations for computer vision and number one in AI and machine learning in Australia. PhD projects span AIML research themes: Machine learning theory, Robotic Vision, Trusted Autonomous Systems, Medical Machine Learning, Vision and Language Methods and Advance Reasoning and Logic.  Projects with a translational focus include priority areas such as health, space, defence, and agriculture.   

    Supervisor(s)TBA
    Contact: Professor Anton van den Hengel; anton.vandenhengel@adelaide.edu.au

  • Other research related to information capability, space or defence