Adaptive PID control of multi-DOF industrial robot based on neural network
Jiyue Wang,Yonggang Zhu,Renlong Qi,Xigui Zheng,Wei Li
DOI: https://doi.org/10.1007/s12652-020-01693-w
IF: 3.662
2020-01-08
Journal of Ambient Intelligence and Humanized Computing
Abstract:The control system of parallel robot, especially industrial robot, is a very complex multi-modal nonlinear system, which has the characteristics of time-varying, strong coupling and strong nonlinearity. Trajectory tracking control algorithm is a very important part of industrial robot control system. It is required that the algorithm can realize the continuous tracking of each joint of the robot and the processing and tracking of the desired trajectory. However, due to the strong influence of acceleration and speed on the trajectory tracking of industrial robots, the corresponding control difficulty and control accuracy are seriously affected. Based on the core idea of fuzzy neural network algorithm, the functional relationship between control error and arrival degree is established to improve the control quality of industrial robots. At the same time, combining with the PID feedforward control algorithm, the self-adaptive adjustment of PID parameters is realized, and the accuracy of the tracking algorithm is improved. In order to verify the superiority of the proposed trajectory tracking control algorithm over the traditional PID algorithm, the model of industrial robot is established by using the virtual simulation system Adams. At the same time, the model is simulated by joint experiment. Experimental results show that the trajectory control algorithm based on the proposed trajectory control algorithm is effective. This method has good control accuracy and stability.
computer science, information systems,telecommunications, artificial intelligence