Underwater Target Signal Classification Using the Hybrid Routing Neural Network

Area of Science:

• Signal processing and analysis
• Artificial intelligence in acoustics

Background:

• Underwater target recognition (UTR) is crucial for signal analysis.
• Conventional methods struggle with rapid and accurate UTR in challenging underwater acoustic conditions.

Purpose of the Study:

• To develop a novel deep learning method for improved underwater target recognition.
• To address the limitations of conventional approaches in classifying underwater acoustic signals.

Main Methods:

• A novel deep learning network (DLN) featuring a hybrid routing structure was designed.
• The network incorporates multiple routing structures and auxiliary branches for feature exchange.
• Time-domain signal features are abstracted and processed within the network.

Main Results:

• The proposed deep learning method demonstrated superior performance in underwater signal classification.
• The hybrid routing network effectively abstracted and utilized signal features.
• Experimental results confirmed the advantages of the new network over conventional techniques.

Conclusions:

• The novel hybrid routing deep learning network offers significant advantages for underwater signal classification.
• This approach provides a more effective solution for underwater target recognition challenges.
• The method shows promise for advancing the field of underwater acoustic signal processing.