Research Of Predictive Control Technology Based On Neural Network Implement To Robot

Up to now, intelligent control has been applied widely in the uncertain and nonlinear system. In this thesis, we aimed at the application of the fuzzy and neural network and predictive control technology in the motion control and try to design a good control system for the robot.The theory of the fuzzy logic and neural network was introduced and the combinations of the fuzzy and neural network were expounded. The fuzzy control, neural control and fuzzy neural control for this vehicle were provided. The paper introduced the traditional way which was used to calculate the controller's parameters. Also analyzed some restricts of this way. Through an exalmple it pointed out some conditions which must to fit when to use this way successfully.The paper discussed the way how traditional predictive control construct their inside model and pointed out the weaknesses of these ways. It put forward a new method to get the controller's parameters.Due to the fuzzy neural network's merit which has compact structure and self-adjustment ability here we implement this arithmetic based on the fuzzy neural network and the paper provided a particular course to design this controller.The computer simulation result together with the experiment to a locomotive robot showed that arithmetic works. After following the experiment's process and analyzing the results, it also pointed out the shortcoming of this arithmetic and took a outlook to somesignificative jobs.