Direct method for model reference adaptive PI controller using the gradient approach

Two types of PI controller configurations considered in this paper for the direct method of model reference adaptive PI controller using the gradient approach. The authors have generated significant insight on the inner workings of the gradient method from a 3-D spatial perspective. The type of PI control rules and the corresponding adaptation rate /spl theta/~/spl dot/ generate certain adverse effects on the overall adaptation trend. Extensive results of the adaptive algorithms for both the standard textbook PI and the modified PI control rules are presented for discussion.<>