Polynomial Methods in Optimal Control and Filtering

This book aims to demonstrate the power and breadth of polynomial methods in control and filtering. Direct polynomial methods have previously received little attention compared with the alternative Wiener-Hopf transfer-function method and the statespace methods which rely on Riccati equations.

The book provides a broad coverage of the polynomial equation approach in a range of linear control and filtering problems. The principal feature of the approach is the description of systems in fractional form using transfer functions. This representation leads quite naturally and directly to the parameterisation of all 'acceptable' feedback controllers for a given problem in the form of a Diophantine equation over polynomials. In the polynomial equation approach, this direct parameterisation is explicitly carried through to the synthesis of controllers and filters and, further, to the computer implementation of numerical algorithms.

The book is likely to be of interest to students, researchers and engineers with some control and systems theory or signal processing background. It could be used as the basis of a graduate-level course in optimal control and filtering. The book proceeds from the necessary background material presented at a tutorial level, through recent theoretical and practical developments, to a detailed presentation of numerical algorithms.

Kenneth Hunt is the holder of a Royal Society of Edinburgh Personal Research Fellowship and has recently joined the research institute of Daimler-Benz AG in Berlin. He obtained a BSc in Electrical Engineering in 1984 and a PhD in Control Theory in 1987, both from the University of Strathclyde in Glasgow, Scotland. From 1987 to 1989 he was with BBN Systems and Technologies and then for two years was a member of the Control Group in the Department of Mechanical Engineering at the University of Glasgow.

His current research areas include adaptive control, the polynomial equation approach to optimal control and the application of connectionist architectures to nonlinear control problems. Recent forays into Al have produced interesting results in control applications to machine induction. Currently he is applying genetic algorithms to control system optimisation and system identification.

Along with being the author of numerous journal and conference papers, Kenneth Hunt is co-editor of Neural Network Applications in Control (Peter Peregrinus Ltd, 1995).