非線形メカトロサーボ系におけるロバストなファジィ・ニューラルネットワークベースト制御 : 第1報,シミュレーションおよび実験による検証(機械力学,計測,自動制御)

抄録

This paper proposes a novel robust fuzzy-neural network based control for the mechatoronic positioning servo systems that have nonlinear characteristics such as friction, backlash, variations of load and system parameters, and unknown disturbances. The proposed fuzzy-neural network based control scheme consists of three elements : a) a feedback compensation of velocity and position : b) a feedforward compensation which has inverse dynamics of the PD controlled plant based on neural networks ; and c) a nonlinear deviation compensator based on fuzzy neural networks. Computational and experimental results for one degree-of-freedom positioning system are shown to confirm the validity of the proposed controller.