James Marshall
James Marshall is Professor of Theoretical and Computational Biology at the University of Sheffield, where he is head of the Complex Systems Modelling Group in the Department of Computer Science, an affiliate member of the Evolution and Behaviour Research Group in the Department of Animal and Plant Sciences, and a member of Sheffield Robotics. After obtaining a Bachelors degree in Computer Science James combined doctoral studies into the evolution of cooperation with a career developing AI technology for computer games at Sony. He went on to postdoctoral positions at Imperial College London and the University of Bristol, then academic positions at Bristol and Sheffield. James’ interests are predominantly in animal behaviour, its evolution, and the interplay of these topics with engineering. He is the author of ‘Social Evolution and Inclusive Fitness: An Introduction’ (Princeton, 2015). His research into collective robotics is currently funded by the European Research Council, and his work on flying robots by a Programme Grant from the Engineering and Physical Sciences Research Council.
Tools for state-of-the-art collective robotics
Collective robotics has frequently been limited by a number of factors particularly, but not limited to, the scale and sophistication of experiments that can be conducted, and the rigour of modelling analyses. The scale of experiments has been addressed by the development of highly scalable collective robotics platforms, but this necessarily correlates with a physical simplicity that limits the kinds of experiments than can be performed. Similarly, advanced analytic techniques have been imported from disciplines such as statistical physics, to analyse models, but these require significant expertise and training to apply. In this talk I will present two tools developed as part of the ERC Distributed Optimal Decision Making Algorithms (DiODe) project. The first, Augmented Reality for Kilobots (ARK), aims to provide a highly flexible, robust architecture for conducting sophisticated ‘mixed reality’ experiments with a highly scalable robot platform. The second, Multiscale Modelling Tool (MuMoT), aims to make sophisticated analytic techniques available for practitioners to apply to their models with the minimum of prior experience, thereby standardising such analysis in the field.
Credit: Salah Talamali