Knowledge-based extension of model-referenced adaptive control with application to an industrial process

Abstract

This paper studies the development, implementation, and performance of three different types of model-referenced adaptive control with respect to a specific practical application. Conventional model-referenced adaptive control is used as the basis of comparison of the performance of two knowledge-based techniques. Fuzzy logic is used to represent the knowledge base of these techniques, and the compositional rule of inference is used for decision making. In one knowledge-based technique, the parameters of a low-level direct digital controller are adapted by the inference engine so that the physical system tracks a reference model, which specifies the desired performance. In the other knowledge-based technique, the reference input to the system is adapted to achieve the desired performance, as specified by a reference model. Both analytical and practical details of the development and implementation of the adaptive control techniques are systematically presented. Development of the knowledge base is described. Representative experimental results from the three techniques, as applied to a prototype fish cutting machine are given. The performance of the techniques is evaluated on the basis of the developmental experience and the experimental results.