From robot swarms to ethical robots: the challenges of verification and validation

Alan Winfield

Bristol Robotics Lab, University of the West of England, Bristol

Abstract

This lecture will be presented in two parts

Part 1 From robot swarms... Swarm robotics is a new approach to multi-robot systems in which control is completely decentralised and distributed. Inspired by the self-organising collective behaviours of social insects each robot has a simple set of behaviours with local sensing and communications. The overall desired system behaviours emerge from the local interactions between robots, and between robots and their environment. The promise of swarm robotics is that it offers the potential of highly adaptable, scalable and resilient systems, able to tolerate the failure of many individual robots without loss of overall system (i.e. swarm) functionality. Significant engineering challenges, however, are firstly how to design the emergent, self-organising properties of the swarm, and secondly how to assure – and ideally verify – safe and dependable operation. Here I will briefly (1) introduce swarm robotics, illustrating the state-of-the-art with respect to current research, both within the Bristol Robotics Lab and elsewhere. (2) Using adaptive swarm foraging as a case study, we will address the engineering challenges of mathematical modelling and optimization. Then (3) we will analyse the robustness and scalability of a swarm robotic system by developing a model of its reliability.

Part 2 ...to ethical robots If robots are to be trusted, especially when interacting with humans, then they will need to be more than just safe. In this talk I will outline the potential of robots capable of modelling and therefore predicting the consequences of both their own actions and the actions of other dynamic actors in their environment. I will show that with the addition of an ‘ethical’ action selection mechanism a robot can sometimes choose actions that compromise its own safety in order to prevent a second robot from coming to harm. An implementation with both e-puck and NAO mobile robots provides a proof of principle by showing that a robot can, in real time, model and act upon the consequences of both its own and another robot’s actions. I argue that this work moves us towards robots that are ethical, as well as safe. I will conclude by discussing the challenges of verification and validation of an ethical robot.

Recommended reading (*key papers)

Part 1:

Winfield AFT and Nembrini J, 'Safety in Numbers: Fault Tolerance in Robot Swarms', Int. J. Modelling Identification and Control, 1 (1), 30-37, 2006.

Liu W, Winfield AFT, Sa J, Chen J and Dou L (2007), 'Towards Energy Optimisation: Emergent Task Allocation in a Swarm of Foraging Robots', Adaptive Behaviour_, _ *15 * (3), 289-305, 2007.

*Winfield AFT, Liu W, Nembrini J and Martinoli A (2008), 'Modelling a Wireless Connected Swarm of Mobile Robots', Swarm Intelligence, 2 (2-4), 241-266, 2008. Final accepted version (pdf).

*Liu W and Winfield AFT (2010), 'A Macroscopic Probabilistic Model for Collective Foraging with Adaptation', International Journal of Robotics Research, 29 (14), 1743-1760, 2010. doi:10.1177/0278364910375139. Final accepted version (pdf).

Dixon C, Winfield A, Fisher M and Zheng C (2012), Towards Temporal Verification of Swarm Robotic Systems, Robotics and Autonomous Systems, 60 (11), 1429-1441. Download pdf (accepted version).

*Bjerknes JD and Winfield AFT (2013) On Fault Tolerance and Scalability of Swarm Robotic Systems, pp 431-444 in Distributed Autonomous Robotic Systems, Springer Tracts in Advanced Robotics, Volume 83, Eds. A Martinoli et al, Springer-Verlag, Springer Berlin Heidelberg. Download pdf (accepted version).

*M. Brambilla, E. Ferrante, M. Birattari, and M. Dorigo (2013). Swarm Robotics: A Review from the Swarm Engineering Perspective. Swarm Intelligence, 7(1):1-41, 2013.

Millard, A. G., Timmis, J. and Winfield, A. F. (2014) Towards exogenous fault detection in swarm robotic systems. In: Natraj, A., Cameron, S., Melhuish, C. and Witkowski, M., eds. (2014) Towards Autonomous Robotic Systems. (8069) Springer Berlin Heidelberg, pp. 429-430.

Part 2:

*Woodman R, Winfield AFT, Harper C and Fraser M (2012), Building Safer Robots: Safety Driven Control, International Journal of Robotics Research, 31 (13), 1603-1626. Download pdf (accepted version).

Winfield, A. F. (2014) Robots with internal models: A route to self-aware and hence safer robots. In: Pitt, J., ed. (2014) The Computer After Me: Awareness And Self-Awareness In Autonomic Systems.1st. London: Imperial College Press, pp. 237-252.

*Winfield, A. F., Blum, C. and Liu, W. (2014) Towards an ethical robot: Internal models, consequences and ethical action selection. In: Mistry, M., Leonardis, Aleš, Witkowski, M. and Melhuish, C., eds. Advances in Autonomous Robotics Systems: Proceedings of the 15th Annual Conference, TAROS 2014, Birmingham, UK, 1-3 September 2014, pp. 85-96.

Jones, S., Studley, M. and Winfield, A. F. (2014) Mobile GPGPU acceleration of embodied robot simulation. In: Headleand, C. J., Teahan, W. J. and Ap Cenydd, L., eds. (2014) Artificial Life and Intelligent Agents. First International Symposium, ALIA 2014, Bangor, UK, November 5-6, 2014. Revised Selected Papers. Springer, pp. 97-109.

*Dennis LA, Fisher M and Winfield AFT (2015), Towards Verifiably Ethical Behaviour, arXiv:1504.03592v1.

General

Winfield AFT, Robotics: A Very Short Introduction, Oxford University Press, ISBN 978-0-19-969598-0, 168pp, 2012.