An Adaptive Fuzzy Self-Tuning Inverse Kinematics Approach for Robot Manipulators

Abstract

In order for a robot manipulator to reach a desired position, an accurate knowledge of kinematics is required. Also, the Jacobian matrix of the robot manipulator should be nonsingular. However, when the robot deals with objects of unknown parameters, the overall kinematics becomes uncertain and changing. Furthermore, the non-singularity of the Jacobian matrix cannot be guaranteed. Fuzzy logic control is a good candidate technique to deal with uncertain kinematics, and Jacobian matrix. Nevertheless, the conventional fuzzy logic control is not adequate to develop a robust and efficient solution for the inverse kinematic problem. In this paper, a new adaptive fuzzy self-tuning control system for robot manipulators is developed.  The developed system proposes two methods for reducing the numbers of rules and number of fuzzy inputs which significantly reduce the computational complexity.  The developed simulations conducted on 2 and 3 DOFs robot manipulators show the effectiveness of the proposed approach.