不依赖对象模型 ,在前馈 -反馈定值控制系统中 ,借助神经网络构成前馈控制器 ,以反馈输出引导网络权值及输出的调整 ,使网络逐步学成前馈补偿功能 ,并最终在控制中占据主导地位 ,实现对主要可测干扰的补偿 .文章分析了神经网络前馈控制器的作用效果 ,并与根据精确模型设计的常规前馈控制器的作用特性进行了比较 .文中采用两种不同方式对神经网络进行训练 ,仿真结果证实了在模型未知的条件下 ,利用神经网络实现前馈控制的有效性 . Without relying on the object model, a feedforward controller is constructed by neural network in the feedforward-feedback set-point control system to feedback the output to adjust the weights and outputs of the network, so that the network gradually becomes a feedforward compensation function. Finally, Control, and realize the compensation of the main measurable disturbance.This paper analyzes the effect of the neural network feedforward controller and compares it with that of the conventional feedforward controller designed according to the exact model.In this paper, two Different ways of training the neural network, the simulation results confirm the effectiveness of the feedforward control using neural network in the unknown model conditions.