Randomness-Enhanced Grey Wolf Optimizer for Inverse Kinematics Solution of Redundant Robotic Manipulators

Abstract

In this study, a randomness-enhanced grey wolf optimizer (REGWO) is proposed to efficiently solve the inverse kinematics (IK) of redundant manipulators. REGWO is an improved version of the grey wolf optimizer (GWO), and it overcomes its shortcomings concerning premature convergence and low population diversity. In the REGWO, more random individuals are introduced to affect population updates, which is beneficial for the global search of the algorithm. A fitness function composed of position and orientation errors is proposed to solve the optimization problem. An eight-degree-of-freedom serial manipulator was applied to verify the superiority of the REGWO. The results were compared with the classical GWO and other four state-of-the-art algorithm results. The simulation results demonstrate that the proposed REGWO can quickly and accurately solve the IK problem of a redundant manipulator.