论文部分内容阅读
讨论了当控制系统出现执行器故障时的一种容错控制方法。该方法中的故障检测器由神经网络构成，而容错控制器则采用控制器重构方式设计，以使故障系统的性能尽可能与正常系统的性能相同。此外，对于具有某类非线性干扰的线性系统，采用了由神经网络构成的非线性控制器，以消除其非线性干扰。仿真结果表明该方法是有效的。
A fault-tolerant control method is discussed when there is an actuator failure in the control system. The fault detector in this method is composed of neural network, while the fault-tolerant controller is designed by controller reconfiguration, so that the performance of the faulty system is as same as the performance of the normal system. In addition, for a linear system with some kind of nonlinear interference, a nonlinear controller composed of neural networks is adopted to eliminate its nonlinear interference. Simulation results show that this method is effective.