Geometric parameter identification of industrial robots based on improved particle swarm algorithm

The common intelligent algorithms, owing to the large number of parameters to be recognized, have the problems of low recognition accuracy, or complex coding, and easily affected by the initial solution. This happens during the process of recognizing the geometric parameters of industrial robots. Based on the second problem mentioned above, this work proposes a new improved algorithm on the basis of particle swarm algorithm. This algorithm first adopts the weight value w of the nonlinear decreasing strategy, so that it can solve this type of high-dimensional function problem of industrial robot geometric parameter identification. Concomitantly, the flock interaction thinking of the flock algorithm has been introduced to the iterative process of the classical particle swarm algorithm, which strengthens the intercommunication of the particle populations. The error model of the industrial robot has been established by d-H method, and then the error model is solved by the improved particle swarm algorithm. The experimental results reveal that the improved PSO algorithm has better optimization accuracy and better stability than the common particle swarm algorithm. The identified results can significantly improve the robot's end localization accuracy.