Dynamics identification of industrial robots via an improved particle swarm optimization algorithm

To meet the demand for high speed and precision in industrial robots, higher accuracy in dynamic modeling is required. An improved particle swarm optimization (PSO) algorithm is proposed for accurate identification of robotic dynamic parameters, and its effectiveness is experimentally validated. First, a dynamic model of the robotic system is constructed using the Newton–Euler recursive method. Next, the excitation trajectory is optimized using the improved PSO algorithm, with the objective of minimizing the condition number of the observation matrix. The dynamic parameters are estimated using the least squares method to obtain the minimal parameter set. Then, the identified dynamic model is validated. Finally, experiments are conducted on a Moka industrial robot to validate the proposed parameter identification method. Experimental results show that, compared to the traditional genetic algorithm, the improved PSO algorithm reduces the condition number of the observation matrix and demonstrates superior performance.