Direct model reference adaptive control: Progress, problems, and prospects

In this paper an attempt is made to describe the progress to date, the problems, and the prospects for Direct Model Reference Adaptive Control Systems. The most significant progress has been in the development of the theory. Algorithms now exist which assure stable operation of the closed adaptive loop for both continuous time and discrete time systems. The most difficult problems remaining stem from the somewhat restrictive nature of the assumptions which must be made concerning the controlled system, and the fact that few hard results are available for the performance of the algorithms in the presence of measurement noise. In spite of these problems, the method has been successfully applied to some relatively simple systems. This fact, combined with recent progress in the theory, gives hope that the prospect of applying it to more complex systems is good.