Optimal digital redesign of hybrid cascaded input-delay systems under state and control constraints

Abstract

This paper presents a new multi-objective evolutionary programming (MOEP)-based digital redesign scheme to find an optimal low-gain digital tracking controller for hybrid control of multi-input, multi-output (MIMO) cascaded analog systems with input time delays, having the constraints on the states and inputs. The proposed global optimization searching technique enables one to find a high-gain optimal analog controller for analog control of MIMO analog systems with specified performances and constrained states. Then, a new MOEP-based digital redesign method, taking into account the inter-sample behavior of the digitally controlled state, is developed to determine the longest possible sampling period under the state and input constraints and to reconstruct the low-gain pulse amplitude modulated (PAM) digital controller for hybrid control of the cascaded input-delay systems. The resulting controlled hybrid system is able to satisfy the state and actuator constraints without significantly affecting the overall performance of the hybrid cascaded systems with input time delays.