ROBUST ADAPTIVE FUZZY CMAC CONTROL FOR UNKNOWN SYSTEMS

This work presents an integrated robust adaptive control scheme that merges the fuzzy control algorithm with the Cerebellar Model Arithmetic Control (CMAC) for unknown systems. The presented adaptation mechanism is used to tune the weight parameters in the CMAC, such that a given ideal stable controller will be best approximated without prior off-line learning phase required. A robust controller is appended to compensate the approximation error of fuzzy CMAC for improving the robustness. Based on the Lyapunov stability analysis the tracking stability can be guaranteed. Demonstrative examples show that the performance of the proposed control schemes is satisfied.