Two-level optimization strategy for fuzzy control design

A two-level optimization strategy is proposed for fuzzy control design. The phase-plane of the dynamic system is firstly divided into two-level nonlinear regions according to response analysis, then a neuro-fuzzy hybrid network is built to optimize the region level by level. By the former, the control process is decomposed properly, quick dynamic response and a stable steady state with high accuracy are achieved coordinately. While through the later, unreasonable factors in the design of the fuzzy control are decreased/eliminated and the effect of fuzzy control is improved by learning. It offers a general, simple and efficient method for fuzzy control design. The optimized fuzzy controller is adopted in an electro-hydraulic servo system, and satisfactory performance is obtained.