On-line gain adjustment of a robot controller, using fuzzy logic.

摘要

A new method for on-line tuning the gains of classical controllers using fuzzy logic is propose.d The design of the fuzzy component of the controller is based on the knowledge of a classical controller's behavior. In the desing of the fuzzy logic controller, an off-line neural network learning process and the human expertise are combined. The designed fuzzy-PD controller is compared with the classical PD and its found that the fuzzy-PD controller saves energy without losing in performance. A two rotational degree of freedom robotic arm is used to demonstrate the effectiveness of the proposed control law.