On line identification for time-varying systems

In this paper, the author applies the local approximation scheme in his previous work on discrete-time adaptive control for deterministic time-varying systems[1] in developing an on line identification algorithm. The algorithm is able to identify unknown systems with quite rapidly time-varying parameters effectively and can be viewed as a generalization of the previous work of Xie and Evans[2]. The parameter estimates minimize a least squares index. Several numerical results of the algorithm are given.