Research On Neural Network Control Systems Based On Coupling Compensator And Their FPGA Implementation

In this Paper,The basic theories about the mutivariable control systems and adaptive control systems. Several new control algorithms based on neural network are studied.The main contents include:It researches the MIMO(multiple input multiple output) control system based on PIDNN(Proportional-Integral-Differential neural network). Traditional decoupling control algorithms have critical disadvantage that they need the exact expression of coupling function in the system; that is, it is necessary to identify the coupling function. For the reason,I studied the control algorithm that can avoid identification of the coupling function. Formula of the algorithm is detaily deduced, simulations about a number of examples are performed by MATLAB software. The simulation results show that we can use the algorithm to complete design of a control system and obtain good system performance indexes under the condition we do not know the coupling function. The paper also describes application of FPGA to realize the coupling compensator in this control system.It studies three kinds of self-adaptive control system based on neural network, that are self-tuning PID control based on the least square method, self-adaptive control based on RBF neural network,the neuron adaptive PID control based on RBF identification. This paper describes the principle of the three kinds of self-adaptive control algorithm, and the simulation results calculated by MATLAB software. It is considered that the algorithms can be applied to single variable nonlinear and time-varying control system, and can achieve good control performance. The paper also uses FPGA to implement three kinds of controllers.There is obvious defect in traditional decoupling control methods in which the precise coupling relationship between the control loops must be known when designing the decoupler, that is say, coupling function should be identified. However, it is very difficult to identify most coupling functions in multi-variable systems. So it is difficult to apply the traditional decoupling control methods to practical systems. In this paper, we create MIMO control algorithm based on PIDNN that can solve the problem,which plagued the academic area for years.A FPGA is a new type of device with the advantages. it has high running speed, wide adaptation, and easy to proggram in it. Compared with other microcomputers such as singlechip, FPGA is more progressive. The paper explores to develop the actual prototypes of controllers and coupling compendators based on our algorithm. And I use the prototypes to verify the control algorithm.The control algorithms and the ways of design FPGA controller, coupling compensator in my paper, can be applied in the chemical industry,oil refining industry, pharmaceutical industry, thermal power generation plant,etc.The algorithms have good application prospects and can produce economic benefits. Meanwhile, reseach work in this paper has some theoretical value.