Stability and performance of an adaptive controller

This paper is concerned with demonstrating techniques for investigating the stability of Model Reference Adaptive Control Systems based on the MIT Rule. It shows that conventional linearization provides insights in to the stability thus confirming the unexpected results obtained in the practical applications. It shows that the overall MIT Rule based system is neutrally stable rather than asymptotically stable. The effect of the adaptation gain on the controller parameter convergence is shown and an experimental method for predicting good values of the adaptation gains is presented.