A hybrid CHAOS-PSO algorithm for dimensional synthesis of a redundant manipulator based on tracking trajectories without or with singularities

Abstract

The paper presents a novel hybrid CHAOS-PSO algorithm for dimensional syntheses of redundant manipulators while not following common design principles of arm length or arm length ratio of a human in previous literatures. More specifically, we employ the clamping weighted least-norm (CWLN) method to solve inverse kinematics of the redundant manipulator tracking desired nonsingular or singular trajectories considering joint limits. Thus, an optimal solution of flexible D-H parameters of manipulator is explored in the search space based on evaluations by the defined fitness functions with the proposed algorithm. As an optimization tool in this work, the hybrid CHAOS-PSO algorithm incorporates chaos into the particle swarm optimization (PSO) algorithm as population initialization and global perturbation technique in order to enhance swarm diversity and escape from premature convergence. To demonstrate validity and practicability of the proposed algorithm compared to PSO, simulations of dimensional syntheses of a 7-degree-of-freedom (7-DOF) redundant manipulator with complex structures along the planed nonsingular or singular trajectories are performed.