A New Error Model for Discrete Systems and Its Application to Adaptive Identification and Control

Abstract : A new error model for discrete adaptive systems is presented in this paper. By introducing a feedback term in the corresponding error model for the continuous case it is found possible to retain the same form for the adaptive laws. It is demonstrated that the output error tends to zero whether or not the input vector is bounded. The new error model has wide applications in both adaptive identification and adaptive control. Two applications are outlined towards the end of the paper to indicate how the error model is derived and the corresponding adaptive laws generated. (Author)