Genetic Algorithm-Based Autonomous Design of Weighting Functions for H_∞ Compensator

Abstract

H_∞ robust controllers are mathematically designed in the frequency domain using loop-shaping techniques under the given control characteristics. In the design process, however, knowledge and/or experience is generally required to define the appropriate weighting functions which satisfy the desired control performance. This prevents the controller design from being simple and cost effective. The paper proposes a novel autonomous design strategy for the weighting functions in the H_∞ compensator design, where GA can optimize both structuring and parameterization of the weighting functions in order to achieve the specified control performance in the time domain. The effectiveness of the proposed design was verified by numerical simulations and experiments using a prototype.