Novel negative-definiteness conditions on the quadratic polynomial function with application to stability analysis of continuous time-varying delay systems

Abstract

When analyzing the stability of time-varying delay systems in view of the Lyapunov–Krasovskii functional, a quadratic polynomial function with regard to time-varying delay is always generated. And it is particularly crucial to determine the negativeness of the matrix of such a quadratic form function for obtaining an analysis result expressed in linear matrix inequalities. This paper proposes a method of tangent intersection in the delay interval segmentation, producing the generalized quadratic convex conditions by further utilizing the cross point between every two tangent lines. It reduces the conservatism of the existing conditions remarkably without requiring unexplainable adjustable parameters and additional decision variables. Benefiting from the newly proposed quadratic convex conditions, the novel stability conditions are derived, the superiority of which is demonstrated through several widely used numerical instances and single area power system PI control example.