An Iterative Dynamic Parameter Identification Method Based on Minimum Parameter Set for Collaborative Robots

To enhance the control performance and human-robot interaction capability of Collaborative robots (Cobots), it is crucial to develop an effective approach to accurately identify their dynamic parameters. Utilizing the global Product-of-Exponentials (POE) formula, Lagrange’s equations and Lie bracket operation, a dynamic formulation approach is employed for Cobots. By converting the dynamic model into a linear form regarding dynamic parameters, a symbolic regression matrix can be obtained. According to the regression matrix, an effective numerical approach based on the QR decomposition technique is proposed to determine the minimum set of dynamic parameters. Based on the regression matrix and the minimum set of dynamic parameters, an iterative Least Squares Method (LSM) is employed to identify the dynamic parameters. Additionally, to improve the robustness of the identification algorithm, an optimization model is established to obtain an excitation trajectory. The effectiveness of the proposed dynamic parameter identification method is validated through simulations on a 7 Degree of Freedom (DOF) collaborative robot with and without measurement noise. In the presence of measurement noise, the average joint torque error is 0.0238 Nm.