T-IK: An Efficient Multi-Objective Evolutionary Algorithm for Analytical Inverse Kinematics of Redundant Manipulator
Di Wu, Guowei Hou, Wenjie Qiu, Bin Xie
Abstract
This letter proposes a new method combined by the parameterization method and T-IK to solve the inverse kinematics problem of redundant manipulators in the position domain. T-IK is an improved multi-objective optimization algorithm based on NSGA-II. By adding population migration strategy and adaptive interval search operator in algorithm, we can maintain the global search ability of T-IK and greatly strengthen its local search ability. This method was applied to an 8-DOF tunnel shotcrete robot whose inverse kinematics algorithm needed to meet the accuracy, continuity and real-time requirement and avoid joint limits. We compared T-IK with Bio-IK, TRAC-IK, NSGA-II, MOEA/D, and ISGABT algorithms on tunnel trajectory, polyline trajectory, and arc trajectory to test its performance. Although T-IK is slightly inferior to MOEA/D in running time, it is much better than other algorithms in almost all indexes.