Adaptive control with exciting signals

摘要

This paper sums up investigations on the identification method with exciting signals. In particular, for a linear plant acted on by additive bounded noise conditions consistency of estimates of the unknown coefficients obtained by this method are given. Notice that such noise is not of necessity random and does not necessarily possess useful statistic properties providing a means for traditional mathematical statistic methods such as the method of least squares, the likelihood maximum method, etc.. The degree of convergence of estimates is refined, as well as the asymptotic normality of estimates is established. The possiblities of the method are exemplified by the adaptive minimax control problem.