Wireless Communication Jamming Recognition Based on Lightweight Residual Network
Xiaole Yang, Tianhui Fu, Yongbin Wang · 2020 IEEE 3rd International Conference on Electronics Technology (ICET) · 2020
Jamming recognition is an important part of communication anti-jamming. To precisely identify various wireless jamming, a lightweight residual network model was proposed. Firstly, signals received were transformed into time frequency images through Short Time Fourier Transform. Then, inputting time frequency images into recognition model, the lightweight residual network based on dilated kernel takes the images as input and directly outputs recognition results, achieving an "end-to-end" identification. The model was able to achieve an 81.21% accuracy even when JSR is -9dB. Compared with common method with manually-designed features and convolutional neural network with plain kernels, the lightweight residual network utilizing the dilated kernel effectively improve performance in recognition with a few of extra computational complexity added.