组合故障的无人机神经网络自适应容错控制

Adaptive Fault-tolerant Control of UAV Neural Network with Integrated Faults

针对四旋翼无人机在传感器偏差故障、双执行器效率损失故障和外界干扰下的容错控制问题进行研究。首先建立组合故障下的四旋翼无人机动力学模型;其次将神经网络融入到观测器的设计,构建一种新的基于神经网络的扩张状态观测器来估计不同故障类型;然后结合自适应理论来设计位置和姿态控制器,并通过权值函数来调节控制器的控制律参数;最后通过李雅普诺夫理论证明了系统的稳定性,数值仿真验证了所提出设计方法的有效性,使得无人机在组合故障下具有更好的容错控制能力。

The fault-tolerant control problems of quadrotor UAV under integrated faults,including sensor deviation fault,dual actuator efficiency loss fault and external interference are studied.Firstly,a dynamic model of a quadrotor UAV under the integrated faults is established.Then,a neural network is integrated into the design of the observer to construct a new extended state observer based on neural network to estimate different fault types.Secondly,dual position and attitude controllers are designed with the adaptive theory,using the weight function to adjust the adaptive law parameters of the controllers.Finally,the Lyapunov theory is used to prove stability of the system,and effectiveness of the proposed design method is verified by numerical simulation,so that the UAV has better fault-tolerant control ability under integrated faults.