Covariance-assignment based control performance assessment

Abstract

The problem of control variance performance assessment with dynamic behavior constraint is investigated based on covariance-assignment theory for multivariable systems. The suboptimal variance benchmark is calculated by the proposed iterative linear matrix inequality (ILMI) technique, taking the dynamic performance of the closed-loop system into account. Thus the resulting variance benchmark is more practical than the conventional single variance benchmark. Consequently, a modified assessing strategy is developed to evaluate the output variance and transient behavior simultaneously. Simulation examples illustrated the validity of the proposed method.