Constrained Variance Design using State Covariance Assignment

Performance objectives which are expressed as upper bounds on the variances of the outputs and inputs are common in stochastic control problems. Unfortunately, all of the control methods which are currently used to attempt to satisfy these constraints deal with the actual objectives indirectly. The purpose of this paper is to provide an alternative methodology which deals more directly with the actual performance objectives. The methodology is based on the theory of state covariance assignment.