Hybrid Attention Module and Transformer Based Fuze DRFM Jamming Signal Recognition
Jikai Yang, Zhiquan Bai, Zhaoxia Xian, Hongwu Xiang, Jingxin Li, Huili Hu, Jian Dai, Xinhong Hao
Abstract
The fuze system is usually affected by jamming signals, especially the digital radio frequency memory (DRFM) based jamming signals. In this letter, we propose a recognition method for the fuze DRFM jamming signals based on hybrid attention module (HAM) and Transformer. Specifically, we first build a backbone network with the combination of the convolutional neural network (CNN) and the Transformer that can better extract the global features of the feature maps. To bridge the CNN and the Transformer, we design a convolutional embedding module (CEM). Moreover, a lightweight HAM is utilized to overcome the missing of the position information in the Transformer and the high complexity of the method. Simulation results show that the proposed recognition method achieves a better trade-off between accuracy and complexity.