ABSTRACT
This chapter examines the technique of controlling linear discrete time plants in the presence of random disturbances employing noise-contaminated measurements of variables that are not necessarily the state variables. It shows that the optimal control policy is to feed back estimates of the state variables. The chapter considers the problem of control of randomly disturbed systems when the measurements of the system state variables are complete and noiseless. It aims to evaluate the average performance of the randomly disturbed system along with the ability of the feedback control to limit random fluctuations in the state variable sequence. The chapter explains the expressions necessary to predict the statistical average performance of the resulting control system. It investigates the resulting closed-loop system dynamics by examining the closed-loop pole locations given by the roots of relation, which is the closed-loop characteristic equation.
