Research on GNSS Jamming Signal Recognition Based on Improved-YOLOv8n Model

With the increase in GNSS jamming activities, accurately identifying the type of jamming signal has become the top priority. In this paper, a spectrogram dataset containing no jamming signal and five common types of jamming signals are created, and the YOLOv8n model is used for recognition, training, and validation. The YOLOv8n-C2fSA model, which is based on the Shuffle Attention, is introduced to enhance the recognition accuracy. Simulation results show that the accuracy of YOLOv8n and YOLOv8n-C2fSA reaches 97.9% and 98.9%, respectively. Considering the accuracy and computational complexity of the model, the comprehensive performance of YOLOv8n-C2fSA is better than other improved models, and it also performs well on the public dataset.