Performance limits analysis of linear predictive control systems with input saturation

Abstract

This paper proposes a novel method to analyze performance features of linear model predictive control (MPC) systems with input saturation. The performance of MPC is compared with that of saturated control to find out how well MPC can perform under input saturation case. By exploring the properties of Riccati difference equation (RDE) and geometric structure of linear MPC, upper and lower bounds of the ratio between performance of MPC and saturated control are obtained. The establishment of such relationship provides a feasible approach to quantify the performance enhancement by introducing receding horizon optimization method into input saturated systems.