Study on Frequency Hopping Signal Detection and Identification Based on YoLov3

Frequency-hopping communication is extensively utilized in military and civilian communications owing to its advantages such as good anti-interference, low interception probability, and strong networking capability, but there are some difficulties in its detection and identification. Based on this, a frequency hopping signals modulation detection and recognition approach on the basis of YoLov3 is put forward in this study. The method converts the problem of frequency hopping signals modulation detection and recognition to diagram targets detection that has been completely addressed using available deep learning algorithms such as YoLov3. Firstly, the time-frequency transformation of frequency hopping signals is carried out by a short-time Fourier transform, and the frequency hopping signals convert to different diagrams. Then the signal images are detected and recognized based on YoLov3. The simulation outcome demonstrates that the method presents a good recognition accuracy for the detection and recognition of frequency hopping signal modulation.