Human-Inspired Machining Robot Posture Optimization via Multi-Redundant DOFs

Abstract

Robotic machining is gaining increasing prominence in industrial manufacturing, but it still faces considerable challenges in machining feasibility and efficiency of large irregular components (LICs). Most of the existing studies attempt to optimize the robot posture for desired manipulability and stiffness, however, the involved methods typically feature a limited number of optimizable parameters, resulting in suboptimal effects that fail to meet the machining demands. To address this problem, inspired by human operation practices, we propose a novel posture optimization method that incorporates multi-redundant degrees of freedom (DOFs) for LICs machining robot. Optimization parameters for multiredundant DOFs of both workpiece and tool are determined by mimicking human machining of LICs at first. Next, the posture optimization model is established by combining the workpiece placement, tool collision constraints, and machining task objective function. Then, a novel swallowing fish swarm algorithm is designed to solve the multi-layered complex optimization problem. Finally, series of experiments on robotic machining of automotive flywheel shell verify that the proposed posture optimization method combines the high flexibility of human machining with the consistency of robotic machining, yielding superior machining results.