基于位形相似性聚类的机器人参数标定研究OA北大核心CSTPCD

Based on the similarity mapping between positioning errors caused by non-geometric parameter errors and robot configurations,the robot parameter calibration configuration space was di-vided into multiple subspaces by clustering algorithm,then the parameter calibration was completed in each subspace.The configurations to be compensated were classified into the corresponding sub-spaces by similarity relation,then the robot motions were controlled according to the parameter cali-bration results of the subspaces,and the positioning errors caused by non-geometric parameter errors were compensated effectively.An UR5 robot was used as the calibration object for the experiments.Compared with the ordinary calibration compensation method,the maximum and average of the posi-tioning error mode of the proposed method are reduced by 60.24％and 66.62％respectively after cali-bration compensation.