Robot Measurement Configuration Optimization and Kinematics Calibration Method Based on New Evaluation Index

Abstract

The accuracy of industrial robots is the most serious problem in industrial manufacturing and production. Kinematic calibration can effectively improve the pose accuracy of the robot's end effector. The effect of kinematic calibration is closely related to the optimal configuration selected by parameter identification and the optimization algorithm. To solve the problem of missing information from traditional observable indexes, this paper proposes a new index based on the spatial analysis theory of matrix, which takes the super-parallel polyhedron volume of identification Jacobian as the effective information content. The new index and the condition number are used as the fitness function to run the genetic algorithm. The optimal configuration is found in the workspace of the six-degree-of-freedom parallel robot and the results are compared. After the parameter identification by the least squares method and the regularization method, the spatial position and orientation angle are, respectively, compensated for the end pose error at the discrete point in the whole space. The compensation results show that the error of the optimal configuration based on the new index in the position coordinate is reduced by 84.15% compared with the error before compensation, and the error is reduced by 26.05% compared with the condition number compensation result. The compensation effect is better at the edge position of the space, and the compensation result of the orientation angle is slightly better than the condition number compensation result.