Research Projects

This page lists the major research projects in which I am currently involved or have been involved. As you will see, a common theme of my research is the use of tools and techniques from information theory, communications theory and statistical signal processing to draw insights into the behaviour of wireless communications networks. If you are a prospective postgraduate student with an interest in this type of work please send me an email or drop by my office.

My research is supported primarily by the Australian Research Council through ARC Discovery and Linkage Projects.

Further information can be obtained by looking at my journal publications.

Easing the Squeeze: Dynamic and Distributed Resource Allocation with Cognitive Radio

The radio spectrum is a scarce and valuable natural resource which is being squeezed by the rapid growth in wireless communications. Cognitive radios make efficient use of radio spectrum by dynamically reusing frequencies. This requires cognitive radios to sense the local environment and to control the interference caused to existing users of the spectrum. In this project we will design novel dynamic and distributed resource allocation algorithms for cognitive radios in order to significantly improve their performance. We will do so using techniques from extreme value theory, game theory and mechanism design and large random matrix theory. This work is supported by an ARC Discovery Project jointly held with Tansu Alpcan from the University of Melbourme, Subhrakanti Dey from Uppsala University in Sweden, and Hazer Inaltekin from Antalya International University in Turkey.

Taming Uncertainty: A Stochastic-Geometric Foundation for Complex Wireless Networks

This project will establish a stochastic-geometric foundation for complex wireless networks. We will harness stochastic geometry and random geometric graphs to analyze and design wireless networks. Our results will also be applicable to more general networks such as small-world and scale-free networks which model many real-life networks. This project is significant in that it will allow us to assess the effect of network geometry on wireless network performance, and to tap into intrinsic properties of many complex networks. The project will lead to novel network control protocols and to a better understanding of distributed processes on complex networks such as cascading failures, decentralized search and the spread of viruses. This work is supported by an ARC Discovery Project jointly held with Hazer Inaltekin from Antalya International University in Turkey.

Gigabit Wireless: Setting the Standard for Tomorrow's Broadband

This project will impact the future international standards for wireless broadband access. The research team will meet the challenge of designing an integrated wireless access technology that offers the maximum possible data rate for mobile users while providing competitive data rates for wireless access to the home. This will be achieved by exploiting a new wireless technology (MIMO-OFDM) that provides a superior air interface to current generation systems. We will enhance this technology with novel approaches to resource allocation and network architecture design that have the potential to drastically increase the system capacity. This work was supported by an ARC Linkage Project jointly held with Stephen Hanly, Subhrakanti Dey and Brian Krongold.

Closing the Gap: Fundamental Capacity Limits for Interfering Wireless Networks and Practical Methods to Get There

The rate, in bits per second, that can be sent across a wireless communication link is severely limited by interference from other wireless links. Surprisingly, there is currently no complete understanding of the fundamental limits to communication rates in the presence of interference. In this project, we will harness very recent progress on the problem of two interfering links, to provide bounds on communication rates when there are many interfering links. Today, it is possible to derive control algorithms to schedule wireless links to avoid strong interference, but without limits, no-one knows how to assess the performance of these algorithms. This project will provide the limits, and also derive novel near-optimal control algorithms. This work was supported by an ARC Discovery Project jointly held with Stephen Hanly and Subhrakanti Dey.

Optimal Deployment of Wireless Sensor Networks

Wireless sensor networks consist of coordinated sensing devices that offer us new ways to understand and interact with the physical world. For a given technology, the key to both optimising the quality of area monitoring and minimising the cost of a sensor network lies in deciding how best to deploy the sensors. We aim to develop powerful new methods to get the best performance from a planned sensor network. This work was supported by an ARC Discovery Project jointly held with Doreen Thomas and Marcus Brazil.

Estimation, Control and Communications

Wireless sensor networks have the potential to dramatically improve the way we do many things, from monitoring the environment and human health to managing complex defence systems. At the heart of all of these monitoring and management tasks lie the fundamental problems of estimation and control. When sensors and controllers are scattered throughout space and share a common communication medium, the estimation and control problems must be tackled taking into account the inherent communication constraints. This project will lead to a deeper understanding of the rich interplay between estimation, control and communications that is integral to the application of wireless sensor networks. This work was supported by an ARC Discovery Project jointly held with Subhrakanti Dey and Girish Nair.

Cooperation and Macrodiversity

A major current focus of my research is on the role of cooperation in wireless networks. Traditional cellular networks are designed so that at any given time, users are connected to one base station only. Each user will cause interference at other base stations (on the uplink) or at other mobile terminals (on the downlink) and this interference limits capacity. A more modern view is to see the network of base stations as a single spatially distributed transmitter and receiver. On the downlink a number of base stations can transmit a signal intended for a particular mobile while on the uplink, the signal transmitted by a mobile is received at a number of base stations. While this idea of base stations cooperating to transmit and receive information is present to a limited extent in existing cellular networks, there is much more to be gained from more extensive exploitation of macrodiversity. We seek to quantify the potential gains from exploiting macrodiversity and to design efficient algorithms for achieving these gains. This work draws upon results from information theory and statistical inference in Bayesian networks and is intimately related to iterative detection and decoding algorithms. This work was supported by an ARC Discovery Project jointly held with Stephen Hanly and Alex Grant.

Multiuser MIMO and OFDM

Advanced communication techniques allow high spectral efficiencies to be achieved over wireless links. Two of the most important techniques revolve around the use of multiple antennas at both transmitters and receivers (MIMO, space-time coding) and the use of many parallel frequency bands for each user (OFDM). While the performance gains of these techniques have been well demonstrated for point-to-point links, the application of these techniques in multiuser environments such as mobile cellular networks is relatively immature. It is crucially important to understand the performance limitations of these advanced communication techniques in multiuser environments and that is the aim of this project. This work was supported by an ARC Discovery Project jointly held with Stephen Hanly and Alex Grant.

Analysis and Design of Multiuser Receivers

In CDMA-based wireless networks, performance measures of interest, such as signal-to-interference ratio, are functions of the signature sequences assigned to each user in the system. This dependence leads to ungainly expressions for network performance indicators and to complex design problems. Recently, a powerful tool has emerged for performance analysis and parameter optimization of CDMA systems. Large system analysis is based on modeling the signature sequences as random quantities and examining the behaviour of the network as the spreading gain and the number of users grows large. In this regime, key performance measures depend on the eigenvalue distributions of large random matrices. Importantly, in many cases of interest, the limiting spectral distributions have a simple form that only depends on the ratio of the number of users to the spreading gain. The end results are performance measures and design rules that are independent of the signature sequences and solely dependent on the key system parameters. In collaboration with colleagues and students, I have employed large system analysis to analyse the performance of multiuser receivers for fading channels and in the design of efficient implementations of linear multiuser detectors.