Current Research

My research interests span neuroethology, computational neuroscience, behaviour based robotics, evolutionary computation and adaptive systems. My approach is best described as computational neuroethology: to combine traditional biology experiments with robotics, modelling and analytical techniques to understand biological systems. Current research topics include: visual navigation in insects and robots, neuromodulation in (real and artificial) neural networks including the INSIGHT project, agent-based modelling in real world problems and analysis of biological imaging data. I also have a general interest in evolutionary computation, machine learning and adaptive systems. 


Visual Navigation in Insects and Robots

In collaboration with the Sussex Insect Navigation Group, I study the acquistion and use of visual memories during navigation in ants and bees, both to understand mechanisms of learning and to develop insect-inspired navigational algorithms for autonomous robots. 

A panoramic video from our new robot as it navigates a route 

With Paul Graham, Bart Baddeley and Phil Husbands, we have developed a novel situated and embodied route navigation algorithm inspired by ants, in which the agent scans the world and moves in the most familiar direction. The algorithm is capable of robust navigation and captures many of the behaviours of ant navigation. We also focus on the role of insects’ innate behaviours – in particular learning walks in ants (with Antoine Wystrach, Alex Dewar and Mike Mangan) and learning flights in bees (with Tom Collett, Natalie Hempel de Ibarra and Olena Riabinina).


Neuromodulation in Neural Networks

Under the supervision of Professors Phil Husbands and Michael O’Shea I studied the diffusible neurotransmitter nitric oxide focussing on its spatio-temporal dynamics with detailed mathematical models of its spread and more abstract Artificial Neural Networks incorporating a virtual neuromodulatory gas, the GasNet. The work inspired analysis of coupled and spatio-temporally embedded networks in general, and through our doctoral student, Dan Bush, we broadened the models to spiking networks. Currently, this work will continue through the INSIGHT project (with Phil Husbands and Kevin Staras) which will study plasticity and learning through MEA experiments and computational modelling.


Agent-Based Modelling and Evolutionary Robotics

I have a broad interest in adaptive systems particularly agent-based modelling and autonomous robotics in which I use evolutionary optimisation. My work on robotics focuses on the adaptivity and evolvability of neural models for controlling autonomous robots. I have also recently started working on agent-based models in social simulations working on the simulation of crowd behaviour with John Drury and Anne Templeton and in human migration with Professor Dominic Kniveton.


Computer Vision and Machine Learning

I use Computer Vision to interpret data from biological experiments. I collaborate with Kate Bentley on image-based analysis of slice data in order to understand vascular growth in cancer and other pathologies. I also use computer vision to track insects from video.  I have expertise in machine learning, especially evolutionary computation and pattern recognition. I am interested in analysing real-world problems in both biology and industry where I focus on hybrid algorithms combining traditional methods with evolutionary optimisation. I have supervised two doctoral projects in this area: Alex Churchill who studied the evolution of tool-sequences in collaboration with Vero and Nick Tomko, in collaboration with Inman Harvey who studied group evolution. I have also provided consultancy in areas ranging from image-based crystallography, drug discovery and financial modelling to Sci-Art projects.