The Study Of X-Y Table Force/Position Neural Network Control

In the recent decades, many efforts have been devoted to the control of X-Ymanipulators. Many researchers try many new ways to force/position controlfrom different point of view. It is known that there exist complex nonlinear,strong coupling and lots of uncertainties in X-Y positioning system, and whenthe end-effector contacts with the environment, the different environmentstiffness have great affection on the system's performance.In this dissertation, parameter perturbation and the uncertainties ofenvironment stiffness are mainly studied based on X-Y positioning tableforce/position control.Firstly, an improved neural networks is proposed to deal with theuncertainty of load and friction parameters existed in the X-Y positioning table.An adaptive genetic algorithm is proposed to update the parameters of the neuralnetwork so as to enhance the control precision and improve the robustness of thesystem.Based on the force/position hybrid control, presenting a method by meansof self-adaptive fuzzy control and CMAC, it can improve the self-adaptivity ofthe control system when the end-effector of X-Y table contact with the workenvironment which has uncertain contact stiffness. A new control strategy ispresented by combining fuzzy-neural control with feedback control underconsidering the uncertainties of X-Y table system based on force/positionhybrid control. Fuzzy-neural network is used to learning the boundary ofenvelope function of uncertainties, and the feedback controller is used toenhance the complete performance of fuzzy-neural control strategy.The theory of hybrid force/position control is clear, however, it is difficultto implement. Then an impedance control strategy with robust performance ispresented aiming at uncertainties of X-Y table, FCMAC is used to learning theuncertainties in order to eliminate disturbance, which have good robust and highvalue in practice.