An Identification Method of Navigation Signal Interference Type based on SqueezeNet Model

With the increasingly complex electromagnetic environment, satellite navigation terminals are often interfered, making it difficult to guarantee service performance. Therefore, the interference monitoring of the navigation signal is urgent. Aiming at the low accuracy of domestic satellite navigation signal interference type recognition, this paper proposes a navigation signal interference type recognition method based on the convolutional neural network: SqueezeNet model. This method first uses the time-frequency analysis of the smooth pseudo Wigner-Ville distribution to transform the timefrequency characteristic domain and generate the timefrequency image. After the time-frequency image is preprocessed, the convolutional neural network of the SqueezeNet model is used to extract the time-frequency image characteristics of the navigation interference signal, so as to realize the automatic, accurate and rapid identification of the common interference signal (Pulse signals, LFM signals, CW signals, BPSK signals) in the satellite navigation field.Finally, through simulation experiments and comparison with other domestic navigation signal interference type identification methods, it is proved that the interference type identification effect of this method is better and the accuracy is higher under the low interference-to-noise ratio.