A Relaxed Mismatch Criterion for Reducing Conservatism in a Lyapunov Stability Analysis

We consider a system whose linear mathematical model includes uncertainties in the system state matrix and, possibly, the input connection matrix A quadratic matching condition is introduced in order to establish a new relaxed mismatch criterion, the satisfaction of which may be used to guarantee quadratic stabilizability via linear state feedback control. We show the relaxed mismatch criterion to generally be less conservative than either the measure of mismatch or the quadratic bound criteria.