On-line PID tuning for engine idle-speed control using continuous action reinforcement learning automata

PID systems are widely used to apply control without the need to obtain a d ynamic model. However, the performance of controllers designed using standa rd on-line tuning methods, such as Ziegler-Nichols, can often be significan tly improved. In this paper the tuning process is automated through the use of continuous action reinforcement learning automata (CARLA). These are us ed to simultaneously tune the parameters of a three term controller on-line to minimise a performance objective. Here the method is demonstrated in th e context of engine idle-speed control; the algorithm is first applied in s imulation on a nominal engine model, and this is followed by a practical st udy using a Ford Zetec engine in a test cell. The CARLA provides marked per formance benefits over a comparable Ziegler-Nichols tuned controller in thi s application. (C) 2000 Elsevier Science Ltd. All rights reserved.