Adaptation in nonstationary applications

In the past decade, stochastic approximation procedures have influenced the design of systems in a variety of system theory applications. These applications are characterized by an uncertainty in the a priori knowledge of the environment in which the system must operate. Howeer, it has been assumed that the environment either is statistically stationary or evolves in a known fashion. In this study a modification of the Robbins-Monro procedure is considered for two classes of unknown nonstationaries. Asymptotic properties of the procedure are considered for both classes of nonstationarities.