基于病态参数分离的机器人运动学标定测量构型优化

For the requirement of a high dexterity robot with high sensitivity of link parameters and precise positioning accuracy, there are the problems of low absolute positioning accuracy, poor parameter identification effectiveness, and calibration robustness in the random measurement configuration of robot kinematic calibration. To address these issues, a robot kinematic calibration measurement configuration stepwise optimization method based on ill-conditioned parameter separation and DETMAX and improved differential evolution (DETMAX-IDE) algorithm is proposed. Firstly, a robot position error model is formulated. Secondly, a comprehensive observability index is developed to evaluate the overall observability and sensitivity for different robot calibration measurement configurations. Finally, ill-conditioned parameters of the robot kinematic position error model are separated. The objective function and constraint conditions are established for optimizing the measurement configuration, the differential evolution algorithm is improved (abbreviated as IDE algorithm), and a step-by-step iterative optimization algorithm based on the DETMAX algorithm and IDE algorithm is presented, which is referred to as DETMAX-improved differential evolution algorithm, and abbreviated as DETMAX-IDE algorithm. The step-by-step iterative optimization of robot kinematic calibration measurement configuration is achieved. Using numerical simulation and experimental robot kinematic calibration, the effectiveness of the proposed method is evaluated. Compared with the random measurement configuration, the experimental results show that the average and the mean square deviation of the robotic absolute positioning accuracy corresponding to the proposed method are improved, with an decrease of 62. 09% and 62. 45%, respectively.