Associate Dean (Research) and Hewlett-Packard Professor of Electronic Engineering

Short Biography

Publications (pdf)

My personal research covers a number of areas:

In this strand I am interested in building mobile robots that may be used as platforms for conducting research into distributed mobile robotics. This work is concerned therefore with software and communications architectures that might facilitate potentially large populations of mobile agents. The LinuxBot, for instance, integrates a commercial-off-the-shelf (COTS) embedded PC-compatible controller, wireless LAN technology and the Linux operating system to produce a remarkably capable (in respect of control and communications) wheeled mobile robot. Here is a page describing the LinuxBot. A new smaller version, the uLinuxBot achieves a similar level of functionality in a wheeled platform just 5cm in diameter and 8cm in height. As part of this work I am interested in ad-hoc wireless networking. In other words, how a potentially large population of identical mobile agents each of which has limited wireless range, can dynamically self-organise to maintain overall connectivity (with no fixed infrastructure).  More recent work is concerned with how we might assure future swarm robotic systems for dependability.

Winfield AFT and Holland OE, 'The application of wireless local area network technology to the control of mobile robots', Journal of Microprocessors and Microsystems (Elsevier), Vol 23/10, pp 597-607, 2000.

This strand of work has much in common with the previous strand (in its use of COTS technology and wireless networking), but is concerned with the control of single vehicles and with the special challenges associated with autonomous operation in very demanding applications, such as Space Robotics.  A current project is concerned with autonomy in the design of planetary landers for small-body (asteroid) landing.

Winfield AFT, 'Wireless Video Tele-operation using Internet Protocols', Proceedings of the 14^th International Conference on Unmanned Air Vehicles, Bristol, April 1999.

Winfield AFT, 'Future directions in tele-operated robotics',  Invited Paper in Telerobotic Applications, ed. T Schilling, Professional Engineering Publishing (IMechE), 2000, pp 147-162.

Goodwin J, Winfield A and Zhu Q, ‘Towards a Generic Architecture for Autonomous Landing Systems’, Proc. Towards Intelligent Mobile Robotics (TIMR 03), Bristol, August 2003.

We have had a number of successful PhD programmes in the area of adaptive neuro-control. This work has focussed on neuro-controllers capable of real-time, on-line learning. In other words, neuro-controllers capable of continuously adapting to the 'unknown' aspects of the plant dynamics. A key aspect of the work has been to underpin the work with rigorous mathematical proofs of stability, because, without guarantees of stability neural control is unlikely to be accepted in safety-critical applications.

More recently, this thread of work in stable control has been extended to a class of behaviour-based controllers.  We believe that this is the first serious attempt to rigorously develop stability proofs for a behaviour-based controller.  Work that could lead to the application of Artificial Intelligence techniques in safety-critical domains, such as aircraft flight controls.

Jin Y, Pipe AG and Winfield A, 'Stable Manipulator Trajectory Control Using Neural Networks', pp 117-151 in Neural Systems for Robotics, eds O Omidvar & P van der Smagt, Academic Press, 1997.

Randall MJ, Winfield AFT, and Pipe AG, 'Stable on-line neural control of systems with closed kinematic chains', IEE Proc-Control Theory Appl., Vol 147, No 6, pp 619-632, November 2000.

Harper C and Winfield A, ‘Direct Lyapunov Design - A Synthesis Procedure for Motor Schema Using a Second-Order Lyapunov Stability Theorem’, Proc. IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS 2002), Lausanne, October 2002.