A Fault Diagnosis Method of Quadrotor UAV Based on Self-attention Mechanism

Due to the different flight actions of the quadrotor unmanned aerial vehicle (UAV), the distribution of condition monitoring data in time dimension is inconsistent. The data-driven diagnostic model is difficult to extract fault features, which reduces the diagnostic performance. In this paper, a fault diagnosis method for quadrotor UAV based on self-attention mechanism is proposed. The attention mechanism is used to extract the long-term dependence of the data in time dimension. Then, a convolution module is added for further feature extraction, which can improve the capability of fault feature extraction. Experimental results show that the accuracy of UAV fault diagnosis under multi-action flight reaches 98.81%, which verify the effectiveness of the proposed method.