Research And Application On Nonlinear Adaptive Inverse Control

Nonlinear adaptive inverse control has been studied in this paper. Adaptive inverse control based on neural network has many advantages such as low requirements of the model, strong adaptive ability and great performance of robustness and fault tolerance. Which can be used in the industrial process control and achieve better control quality. This thesis is to find a control method of nonlinear system with the advantage of rigorous theory and easy-engineered. In the same time, neural network has demonstrated significant superiority to research of nonlinear system. This thesis combined adaptive inverse control method with neural network, putting forward the control method of non-linear adaptive inverse control system.Firstly, this thesis introduced basic theory. Secondly, for the theory of nonlinear adaptive inverse control, the introduction-type fuzzy chaotic neural network (IFCNN) is got via combining chaotic mechanism and fuzzy neural network. BP algorithm used commonly has slow convergence and easy to fall into local minima, but chaotic BP algorithm based on gradient descent searching algorithm has fast convergence to local optimum, which can get out of local minima by relying on chaotic global optimization. Finally converges to the global optimum. So, it can deal with linguistic information, which has adaptive fuzzy reasoning ability, self-learning and chaotic search capabilities and simulate the processing of human brain to process information better. Thirdly, this paper introducing the merits of neural network in nonlinear system modeling and control into nonlinear adaptive inverse control by using approximation ability of neural networks to arbitrary nonlinear function, thus getting forward and backward modeling. Then, constitute a kind of nonlinear adaptive inverse control system based on IFCNN. Also, preliminary discussion and analysis on stability of nonlinear adaptive inverse control system is done. Finally, in order to verify system stability, feasibility and accuracy, the double inverted pendulum is controlled by using nonlinear adaptive inverse control method. Studies have shown that the control structure designed has good robustness and fault tolerance.