基于两阶段卡尔曼滤波的四旋翼无人机自主预测维护

Quadrotor UAV predictive maintenance utilizing a two-stage Kalman filter approach

中文关键词  四旋翼无人机  多执行器退化  健康度  预测性维

英文关键词  quadrotor UAV  multiple actuator degradation  health degree  predictive maintenance

中文摘要

      本文针对多执行器退化的四旋翼无人机, 基于机体性能状态的健康感知, 通过反馈调节权值矩阵实现了无 人机的自主维护. 文中首先建立了四旋翼无人机退化模型, 并基于预测性维护和模型预测控制策略构建了自主预 测维护体系架构; 其次, 采用两阶段卡尔曼滤波法实时估计无人机机体状态和各执行器退化量, 并用熵权法将4个 执行器退化量融合为一综合退化量; 接着, 基于马氏距离健康度计算机体失效阈值, 求解无人机剩余寿命, 在无人机 不满足机体性能和时限性约束时, 依据健康度评价实时调整权值矩阵实现自主维护. 仿真结果表明, 本文提出的自 主预测维护方法能够有效延长机体寿命.

英文摘要

      In this paper, we propose a method to achieve autonomous maintenance of quadcopter unmanned aerial vehicle(UAV) with multi-actuator degradation by adjusting the weight matrix based on the health-aware body performance state. First, we establish a degradation model for quadrotor UAVs and then construct an autonomous predictive maintenance architecture based on the predictive maintenance and model predictive control strategies. Second, a two-stage Kalman filter method is used to estimate the body state of the UAV and the degradation of each actuator in real-time, and an entropy weighting method is used to fuse the degradation of the four actuators into a comprehensive degradation. Then, we calculate the failure threshold of the UAV based on the Mahalanobis distance health degree and solve the remaining lifetime of the UAV. When the UAV does not satisfy the body performance and time-bound constraints, the weight matrix is adjusted in real-time based on the health degree assessment to achieve autonomous maintenance. Simulation results show that the proposed method can effectively extend the life of the airframe.