容错控制系统的一种神经网络控制策略

基于神经网络的函数逼近能力及其容错性，提出了一种神经网络容错控制策略：首先利用系统重构的方法设计控制系统在各种故障情况下的控制律，然后采用一个神经网络来学习这些控制律的特性．学习结束后，将该神经网络作为控制器对系统实施控制．对一个具体的线性系统在传感器故障情况下的神经网络控制进行了仿真研究，结果表明：神经网络控制器能够代替系统原有的控制器．而且在系统发生未知故障时，同样具有容错性．

A Control Strategy with Neural Network for a Fault Tolerant Control System

Based on both the function approximation and fault tolerant abilities of the neural network, a fault tolerant control strategy with neural network is presented. The control rules of the control system in various cases of faults are designed by means of system reconfiguration. Then a neural network is used to learn the features of these rules. When the learning process is over, the neural network will be used as a controller for the system control. Simulation on the neural network control of a practical linear system in sensor fault has been carried out. The result shows that the original controller can be replaced with a neurocontroller, which is fault tolerant when unanticipated faults occur in the system.