GA-based evolutionary design of robust compensator for mechatronic systems

This paper presents an evolutionary algorithm for the robust motion controller design in mechatronic systems, using the genetic algorithm. The motion control system is composed of a robust 2-DOF compensator based on the coprime factorization description. Conventional controller design approaches to the optimization for the compensator free parameters essentially require complicated numerical procedures under the given control specifications. In this study, a simple and practical algorithm for the compensator design for mechatronic systems is proposed. The proposed algorithm is able to autonomously tune the optimal combination of the compensator free parameters to satisfy the specified motion control performance. The effectiveness of the proposed optimal design can be verified by experiments using a prototype, paying attention to the system robustness against variations of the mechanical parameters and the fast convergence of the optimization by the genetic algorithm