Fixed- Time Adaptive Neural Tracking Control for Robot Manipulator with Output Error Constraints

• Published in: 2021 International Conference on Security, Pattern Analysis, and Cybernetics（SPAC pp. 120 - 125
• Main Authors: Huang, Haiqi, Yang, Chenguang, Dai, Shilu
• Format: Conference Proceeding
• Language: English
• Published: IEEE 18.06.2021
• Subjects: Adaptation models; Artificial neural networks; Control theory; fixed-time stability; incremental learning; neural network; output error constraints; robot manipulators; Stability analysis; Trajectory; Transfer functions; Uncertainty
• DOI: 10.1109/SPAC53836.2021.9539941

A fixed-time neural network(NN) adaptive control scheme with output limited function for a kind of general manipulator model with uncertainty is proposed in this paper. Considering that the position of the manipulator in practical applications often needs to meet various constraints, this scheme first converts the problem of seeking a constrained solution into a problem of seeking a stable solution by introducing a transfer function and then designs an output-constrained controller. In addition, the uncertain blocks of manipulator model is estimated by an adaptive NN whose neural nodes are generated dynamically by incremental learning to avoid the decline of approximation performance due to the uncertainty of artificially preset nodes in the traditional network. Finally, the primary controller is designed according to the theory of fixed-time stability, guaranteeing that the system converges in a fixed time as well as the settling time irrelevant to initial state. Simulation experiments verify the availability of the proposed method.