David Clarke has made major contributions to the control, safety and efficiency of industrial processes. The combination of identification and control techniques which led to his generalised least squares method for the optimal estimation of industrial plants with complex, unknown models was among the very first developments of its type in this important field. There are a wide range of applications for his work, ranging from the control of plate glass manufacture and ship stabilisation to business forecasting and physiological control mechanisms.
Further development led to the generalised minimum variance controller. The Clarke–Gawthrop self-tuning algorithm on which this is based has become a benchmark against which all later approaches to this problem have been judged. This work was further complemented by a much used method called generalised predictive control, one of the few classes of control technique which can handle constraints such as quality and safety. He remains vigorous and active in research and is currently working on the wider aspects of safety in control, especially with the development of intelligent self-validating instruments.
Interest and expertise
Engineering, control (incl robotics), Computer engineering (including software)
Adaptive systems, Control systems, Intelligent Process Instruments, Safety of computer control systems