LQG PREDICTIVE OPTIMAL-CONTROL FOR ADAPTIVE APPLICATIONS

The LQG optimal control philosophy is extended so that use may be made of future set-point or reference signal information. The tracking error is defined in terms of an ideal response model so that a model following capability may be introduced. The cost function may also be augmented by the addition of prediction error terms. The controller was derived for adaptive control applications and has many of the desirable features of the new "Predictive Control" philosophy. However, there are fewer variables to choose and guaranteed stability properties are obtained for all weightings or prediction intervals. The self-tuning controller briefly described is almost as simple as existing LQG industrial self-tuners. © 1990.