Intelligent Control Methods With Applications To USM

Intelligent control is a control system to deal with the systems with high uncertainty and complication. It opens a door for controlling complicated problems and has become the research key problem in the control field.Fuzzy control and neural network control are two important intelligent control methods. Fuzzy control is a control method using the concepts of fuzzy set theory, the fuzzy linguistic variable and the fuzzy logic inference knowledge. It simulates the fuzzy thinking of people for the control of some plants or the processes which can not be described by the precise mathematical model. The neural network control is introduced the neural network into the control system to model, control, optimize computation, inference, the failure diagnosis, suitable combination and so on. Fuzzy control and neural network control have been used widely in the fields such as natural science and social science.In this paper, a survey of intelligent theory and intelligent control method is first given; especially the fuzzy logic and neural network theory are introduced in detail. Then, an intelligent method is developed using the combination of fuzzy neural network and simulating human intelligent integral, and is hence applied to the control of ultrasonic motor. The main works of this paper are as follows:1. Developed a new fuzzy membership function form, and guarantee that the membership function has zero point, smooth and the shape variability. Propose a general form of membership function-asymmetrical cosine membership function (Non Symmetric Cosine Membership Functions, NSCMF). It is possible for NSCMF to approach to the Gauss function, asymmetrical Gauss function and Sigmoid function by adjusting the parameters in NSCMF. Therefore, the membership function with NSCMF in the fuzzy neural control system may avoid the difficulties in the membership function selection.2. A simulating human intelligent integral scheme is introduced into the fuzzy neural control process and the stability of control system is increased. Propose a kind of fuzzy neural network controller which has the function of simulating human intelligent integral applying into the online control of dynamic system. This kind of simulating human intelligent integral method can simulate the memory characteristic of people, "memorize" the useful information selectively and "forget" the useless information, and overcome the disadvantages of the conventional integral control, improve the control accuracy.3. Give the structure and the learning algorithm of the intelligent fuzzy neural system. And the numerical experiments are given for controlling a longitudinal oscillation ultrasonic motor. It shows that the control accuracy and control stability are obviously improved by applying the asymmetrical cosine membership function and the neuro-fuzzy control system. The very good control effect are obtained to many different kinds of reference speed by using the control method proposed in this paper, which indicates the controller of this article has the very good adaptive performance.