Exact model matching with disturbance suppression

Abstract

The existing exact model matching (EMM) makes it possible for a plant output to exactly follow the reference model output, if no disturbance is applied to the plant, i.e. EMM has perfect tracking performance. EMM does not however, suppress the disturbance effect positively, and hence, the specification on regulating performance will not be satisfied on certain occasions. Disturbances, as described by any deterministic or stochastic formula, are not considered here; the only assumption is that they are bounded and vary slowly with time. In this paper, a new method to synthesize a disturbance predictor, as well as an EMM design method to suppress the disturbance effect using the prediction, is developed. The same control system is also effective to suppress the plant perturbation such as parameter variation, nonlinearity and unmodelled dynamics.