Optimization Study of Excitation Trajectories for Dynamic Parameter Identification

Abstract

In addressing the issue of identifying dynamic parameters for robotic arms, a trajectory optimization method was proposed using an improved snake optimization algorithm. This method was innovatively built upon the conventional snake optimization algorithm by introducing adaptive adjustment operators in place of fixed coefficients. This adaptation enhanced the global search capability and convergence speed of the snake optimization algorithm. The improved snake optimization algorithm was applied to the optimization design of excitation trajectories in the process of robotic arm dynamic parameter identification. The iterative reweighted least squares algorithm was employed as the parameter identification technique. In the experimental validation phase, a six-degree-of-freedom collaborative robot was chosen as the verification subject. The results demonstrate that, in comparison to conventional excitation trajectory design algorithms, the root mean square deviation of joint torques for the first three joints of the robotic arm decreases by 20.96%, while the root mean square deviation of joint torques for all six joints decreases by 23.58%. This verifies the effectiveness of applying the improved snake optimization algorithm to excitation trajectory optimization design, leading to enhanced accuracy in the dynamic parameter identification.

Keywords

• Collaborative robot
• Dynamic model
• Improved snake optimization algorithm
• Adaptive operator
• Parameter identification