Melbourne School of Engineering
Electrical and Electronic Engineering


This page is in two parts due to my two-year full-time secondment to the Australian Research Council. Only selected publications have been listed below.

Here is a bibtex file containing the bibliographic entries of the papers listed below.



Selected Publications (2009 - 2015)

Differential Geometry in Signal Processing
Stochastic Algorithms

We consider consensus seeking of networked agents on directed graphs where each agent has only noisy measurements of its neighbors' states. Stochastic approximation type algorithms are employed so that the individual states converge both in mean square and almost surely to the same limit. We further generalize the algorithm to networks with random link failures and prove convergence results.

A current hot topic is coordination and consensus in a network, by which is meant that information can be passed from node to node via the links of the network and it is required to find a decentralised algorithm which will allow the nodes to reach agreement about the value of some quantity of interest. (Each node might measure its surrounding temperature and the network must decide in a decentralised way on what the average temperature is, for example.) A rigorous statistical analysis was carried out under quite general assumptions to determine, roughly speaking, how quickly the network could reach consensus.


Ensuring each user experiences a satisfactory quality of service is an important challenge for network designers, especially in wireless networks, where resources are relatively scarce and interference is relatively high. Accordingly, there has been recent interest in bandwidth allocation in wireless ad hoc networks, the focus of this article. After highlighting the main challenges, we survey recently proposed solutions, which address the problem at the network or MAC layer, individually or jointly. We also classify these solutions according to some major design criteria, and suggest the directions of future work on bandwidth allocation.

A rigorous mathematical and information-theoretic analysis was undertaken to determine what the channel capacity (in Shannon's sense) is of a particular channel model commonly used to model a real neuron. It was discovered that the capacity achieving distribution is discrete - a non-obvious and intriguing result. An algorithm exploiting this fact was developed to compute numerically the channel capacity. The results were consistent with experimental evidence.

Selected Publications (Prior to 2006)

Optimization on Manifolds (Differential Geometry in Signal Processing)
Channel Identifiability in Wireless Communications (Algebraic Geometry in Signal Processing)


Affine Precoders and Super-Imposed Training in Wireless Communications
OFDM Systems and Systems with Guard Intervals in Wireless Communications
Filtering of Stochastic Processes
Other Ideas Waiting to be Taken Further
Expository Material