基于神经网络的电动舵回路容错控制方法研究

摘要

In order to restrain the influence of fault and nonlinear in the actuators, a neural network active fault tolerance control method is introduced by applying error feedback learning control law,which is based on neural network model - following adaptive control theory. Considered the dynamics of actuators, the nonlinear mathematical model of actuators is build. Neural network weights are adjusted according to the proposed learning algorithm. The control signal is brought through the neural network learning process which is drove by error signal which is produced by contrasting reference model and real system. Based on the above method, the fault tolerance control systems are achieved by simulation.%为了克服电动舵回路故障和非线性因素的影响,提出一种基于神经网络误差反馈学习的电动舵机容错控制器设计方法,根据电动舵机系统动力学特征建立系统非线性数学模型,采用反馈误差学习控制律进行神经网络容错控制,由参考信号和实际输出信号对比产生的误差信号驱动神经网络学习从而产生控制信号,以达到实现容错重构控制的目的,仿真结果表明,神经容错网络控制器能够达到满意的控制效果.