Fault Diagnosis of AUV Propeller Based on Interference Suppression CLSTM-CNN-Attention

Autonomous underwater vehicles (AUVs) are indispensable equipment in underwater detection, surveying, and investigation, and thruster fault diagnosis is crucial to ensure the safe navigation of AUVs. However, due to the uncertainty of the underwater environment and the unpredictable occurrence of faults, this task is full of challenges. Therefore, we propose a CLSTM-CNN-Attention method for interference suppression for AUV thruster fault diagnosis. Specifically, we introduce a new network structure that combines the improved Long Short-Term Memory Network (LSTM) and Convolutional Neural Network (CNN) in a parallel connection to simultaneously capture the temporal and spatial fault information of the input signal. At the same time, we also introduced a new linear rectification function to improve the model's anti-interference ability. Finally, by adding attention mechanisms to further improve diagnostic performance, the fused information can be better focused on. Experimental results have shown that our proposed method has strong interference suppression ability and outperforms other methods, demonstrating good performance in AUV fault diagnosis.