Adaptive radial basis function neural network bi-quadratic functional optimal control for manipulators

In this paper, a double quadratic optimal functional solution model based on neural network two-stage superposition optimization is proposed to solve the optimal control problem in non-linear manipulator systems where it is difficult to balance the control energy and the proportion of control errors. In the non-linear manipulator control system, the comprehensive optimal control is realized, which uses little control energy to keep the smaller control error. In the model proposed in this paper,firstly, a linear error function is designed to act on the non-linear governing equation, and the uncertainties in the non-linear governing equation are approximated by radial basis function(RBF) network adaptively to form a closed-loop feedback system to realize the optimal control of the non-linear system. Secondly, the parameters to be solved are combined into the solution domain of the bi-quadratic functional, and a new type of recursive neural network is designed to solve the bi-quadratic model with constraints, so as to realize the fast convergence and obtain the solution of the model. The results of theoretical analysis and numerical simulation show that the proposed model can effectively improve the control accuracy, stability, robustness and self-adaptability of the non-linear system, thus realizing the integrated optimal control of the non-linear system. © 2020, Editorial Department of Control Theory & Applications South China University of Technology. All right reserved.