高超声速飞行器自适应神经网络容错控制

针对高超声速飞行器巡航段执行器控制效益损失故障和卡死故障问题,基于高超声速飞行器纵向运动模型,将自适应算法与改进的径向基函数神经网络(RBFNN)方法相结合,设计了一种自适应神经网络容错控制器。所提出的容错控制方法具有无需估计执行器故障值的优点,且设计的控制算法结构简单,无需大量实时计算,可以快速处理故障的发生,确保系统在参数不确定、恒定或时变执行器故障与卡死故障情况下仍具有稳定跟踪能力。最后,仿真验证了该方法的有效性。

Adaptive neural network fault-tolerant control of hypersonic vehicle

Aiming at the problems of actuator loss of effectiveness fault and stuck fault in hypersonic vehicle, this paper proposed an adaptive neural network fault-tolerant control method based on the longitudinal model. The controller was designed by combining an adaptive algorithm with an improved radial basis function neural network(RBFNN) method. It has a simple structure, which can quickly deal with the occurrence of fault and has the advantage of no need to estimate the fault value, thus ensuring that the system has stable tracking capability in the event of parameter uncertainties, constant or time-varying actuator fault and stuck fault. The simulation verified the effectiveness of the controller.