A fuzzy controller with grey prediction for robot motion control

Abstract

The current position error and error change are usually employed in the operation of fuzzy logic controller for picking out an appropriate rule. However, when the system possesses high planning speed or time-varying behaviour, the requisite tracking accuracy is difficult to achieve by adjusting the fuzzy rules. In order to improve control accuracy and system robustness for robotic applications, a fuzzy controller incorporating a system output prediction strategy is proposed to manipulate the robotic motion. The current position error and error change in the fuzzy rules look-up table are substituted in this approach by the predictive position error and error change of the next step, derived from a grey prediction algorithm. This controller was implemented on a four-degree-of-freedom robot. The experimental results show that this fuzzy controller effectively improves the system performance and achieves the envisaged benefit of fuzzy controller implementation.