多变量模糊控制系统的前馈解耦

为实现多变量模糊控制系统的动态解耦，基于前馈解耦思想和神经网络理论，提出了一种多变量模糊控制系统解耦的新方法——模糊前馈解耦法,模糊控制器和解耦部分独立设计，解耦由两层神经网络实现，节点少，其活化函数采用分段线性函数．利用简化的学习算法，根据系统输出误差，在线调整网络权值，从而实现动态解耦而无需辨识被控对象的模型，该方法结构简单且计算量小，适于实时多变量过程控制，仿真证明了该方法的有效性。

Feedforward Decoupling for Multivariable Fuzzy Control System

On the basis of feedforward decoupling and neural network theory, the fuzzy feedforward decoupling method is presented to implement dynamic decoupling control for multivariable fuzzy system. Fuzzy controllers and decoupling units were designed independently, where the neural network with a single hidden layer and a limited number of neurons was employed for decoupling. The activity functions were replaced by segmental linear functions. The simplified training algorithm based on the system output error was used for online adjustment of the network weights, thus realizing dynamic decoupling, which eliminated the need of identifying the plants. This method, with a simple architecture and a small amount of computation is suitable for real-time multivariable process control,and is proved to be effective by simulation results.