An Adaptive State Feedback Controller for a Class of Stochastic Systems

The paper is concerned with the problem of non-zero set-point (NZSP) control of linear stochastic systems with unknown model parameters. It is shown that a convenient solution is possible through the minimization of the state error variance. It is also shown that the proposed combined "NZSP feed-forward and estimated state feedback," control law is simpler to compute than the LQG solution. Further, the new solution has some important advantages over the self-tuning controller. The results are illustrated using a simulated example.