Optimal design of robust vibration suppression controller using genetic algorithms

This paper presents an evolutional compensator design for motion control systems using genetic algorithms (GAs). The control system is composed of a robust two-degrees-of-freedom (2DOF) compensator based on the coprime factorization description. A feedback compensator in the 2DOF control system is theoretically designed under the /spl mu/-Synthesis framework to ensure the robust stability because the real plant mechanism includes structured uncertainties, e.g., the frequency perturbations of vibration modes. On the other hand, a feedforward compensator is optimized by GA paying attention to the robust servo characteristics against the mechanical parameter variations, where the structuring and parameterization of the compensator can be autonomously achieved to satisfy the desired servo characteristic with the resonant vibration suppression performance. The effectiveness of the proposed controller design has been verified by experiments using a prototype.