A combination network of CNN and transformer for interference identification

Area of Science:

• Electronic Warfare and Signal Processing
• Deep Learning for Communication Systems

Background:

• Accurate communication interference identification is vital for effective electronic countermeasures.
• Existing deep learning methods, like CNNs and transformers, often fail to integrate both local and global signal characteristics.
• Distinguishing contextual semantics in 1D signal data presents a significant challenge for current approaches.

Purpose of the Study:

• To develop a novel deep learning network that effectively utilizes both local and global signal features for interference identification.
• To improve the recognition accuracy of communication interference in electronic warfare scenarios.
• To address the limitations of existing methods in handling 1D signal data and capturing time-frequency characteristics.

Main Methods:

• A novel network architecture combining Convolutional Neural Networks (CNNs) for local feature extraction and transformer-based attention mechanisms for global context.
• Utilizing CNNs instead of traditional word embedding to better capture intrinsic features of 1D signal data.
• Integrating a cross-attention mechanism to fuse temporal and spectral domain information, eliminating the need for separate time-frequency analysis.

Main Results:

• The proposed network significantly enhances the utilization of both local and global signal characteristics.
• Experimental results show a substantial improvement in recognition accuracy compared to existing state-of-the-art methods.
• The integrated cross-attention mechanism effectively captures essential time-frequency signal properties.

Conclusions:

• The novel hybrid CNN-transformer network offers a superior approach for communication interference identification.
• The method demonstrates high efficacy in electronic warfare applications by improving signal recognition accuracy.
• This approach provides a more robust and integrated solution for analyzing complex signal data.