Fine-tuning fuzzy control of robots

Abstract

A robotic system including the robot manipulator and motors is highly nonlinear, heavily coupled and multivariable in a non-companion form. To guarantee stability of the tracking control, a multivariable fuzzy controller may be designed; however the controller becomes computationally extensive. In contrast, a PD fuzzy controller is computationally simple in a decentralized structure; however a precise performance may not be provided. To remove the shortcoming, this paper presents a novel fine-tuning fuzzy control of robots which is computationally simple with a guaranteed stability. Its performance is better than the PD fuzzy controller due to having the fine-tuning ability. Unlike the adaptive fuzzy control which adapts all fuzzy rules, the fine-tuning fuzzy control adapts only one fuzzy rule. The convergence analysis is efficiently used to obtain a simple design with high-accuracy response, robust tracking performance and guaranteed stability. Simulation results using an articulated electrically driven robot manipulator shows superiority of the fine-tuning fuzzy control over the PD fuzzy control in the set point and tracking control.