Adaptive time optimal fuzzy control

The author shows how techniques of artificial intelligence and fuzzy reasoning could be used to implement a time optimal control policy when an exact mathematical model of the process is not known. The controller has a simple knowledge base with knowledge about control policy for different starting errors, but it also has adaptive and self-learning properties. After each run the controller adjusts itself using simple metarules in order to improve process response. Theoretical foundations are illustrated by results of laboratory experiments with a two-degree-of-freedom mechanical system.