Research On AUV Self-noise Suppression Technique Based On Informer

Abstract In order to enhance the performance of Autonomous Underwater Vehicles (AUVs) in underwater target detection, this paper proposes an AUV self-noise suppression method based on the Informer model. Initially, the method involves the Informer model extensively learning the historical self-noise of AUVs. Subsequently, based on the acquired knowledge, the Informer model generates predicted AUV self-noise. Finally, by subtracting the predicted self-noise from the signals received by the hydrophone, the suppression of AUV self-noise is achieved. The effectiveness of this method has been demonstrated through simulation experiments, with noise reduction effects of approximately 3 dB.