Intelligent control using fixed and adaptive models

The paper deals with the problem of intelligent control using multiple identification models, switching and tuning. This approach was introduced by the authors (1994) to improve the transient performance of adaptive systems in a stable fashion. The method is extended in this paper to the case where both fixed and adaptive identification models are included. The principal contribution of the paper is the demonstration of the stability of such adaptive switching systems for a general class of switching and tuning schemes. Extensive simulation results reveal that substantial improvement in performance can be obtained with the methods proposed, and this has been experimentally verified by others in robot control problems.<<ETX>>