Adaptive minimum variance control for linear discrete time systems

An algorithm for adaptive minimum variance control of unknown plants in unknown environments is presented. The derivation of the algorithm is based on the use of least-square techniques. Experimental results show that for stationary plants of known order the adaptive controller converges to the true optimal controller. An algorithm that exhibits exponential forgetting is suggested for use in the non-stationary case and simulations show that the performance of the controller approaches very well the optimal performance.