Fuzzy Neural Network Sliding Mode Control For Direct Drive XY Table System

This topic is derived from the National Natural Science Foundation of China(51175349) and Liaoning Province Academic Talents Support Program funding(LR2013006). Nowadays high expectations of industry for CNC feed system increasinglyand the trends of these expectations is three high-tech that the so-called three high-tech arethree-speed, high-precision and high-composite.In this paper, the single axis trackingaccuracy、immunity、contour errors and other issues of two permanent magnet synchronouslinear motor (PMLSM) direct driven XY table system which Vertical movement in thedirection of movement. The fuzzy neural network sliding mode control and iterativelearning cross-coupling control direct drive XY motion system are designed.Firstly, large numbers of relevant articles from all over the world are referenced inthis paper. And the contour control of direct drive XY table driven by permanent magnetlinear synchronous motor development is introduced in this paper. The three main controlmethods are expounded in this paper, the traditional control method, the modern controlmethod，and the intelligent control method.Secondly, in detail introduced the working principle of PMLSM and the mathematicalmodel of the direct drive XY table. And analyzed contour error model of straight line,circle and arbitrary input trajectory for the direct drive XY platform, summarizes thecauses of contour error.Then, on the basis of the above described, in order to suppress external disturbances,system parameters change and unmatched, since the learning ability of RBF neuralnetwork sliding mode control and cross coupling control system have been designed.Using self learning ability of RBF neural network control to train switching function ofsliding mode control, use the neural network output instead of the sliding mode controloutput, to reduce the position tracking error. Using the cross-coupling control to solvetwo-axis motor parameter mismatch problem which is to reduce the contour error. Andusing Matlab/Simulink to simulate and analysis. Finally, in view of the network connection physical meaning is not clear and slidingmode control chattering problem can not be well to improve. Design a fuzzy neuralnetwork sliding mode and iterative learning cross coupling control system. Fuzzy neuralnetworks sliding mode position controller compensates the insufficient for the three kindcontrol methods, at the same time it also has the advantages of the three kinds controlmethod. In biaxial, system related to the periodic contour error which is related to theprocessing cycle cannot be eliminated just by cross-coupling control, iterative learningcross coupling control can solve the problem of periodic error, make the system has bettercontrol effect, reduce the contour error. And using Matlab/Simulink to simulate model andanalysis, and contras the simulation results of the RBF neural network sliding mode andcross-coupling control contour system with the fuzzy neural networks sliding modecontour and iterative learning cross coupling control system, the later has better controleffect and the outline of precision.