Aiming at the problem of low recognition rate of modulation modes of complex multi-class radar signals under low signal-to-noise ratio conditions, this paper proposes a radar signal modulation mode recognition method based on time-frequency analysis and deep learning. The signal time domain waveform is transformed into a two-dimensional time-frequency image using CTFD (Cohen class Time-Frequency Distribution) time-frequency analysis to more clearly characterize the signal characteristics; the time-frequency image is preprocessed by graying and bicubic interpolation operations to reduce the number of image channels and size, thereby reducing the amount of data input to the deep learning model; the number of input and output channels is further adjusted to construct a small EfficientNet network, and then multiple small networks are processed in parallel to construct a split network EfficientNet-B0-Split3, and the time-frequency image is input into the network to realize radar signal modulation mode recognition. Experimental results show that when the signal-to-noise ratio is −8 dB, the overall recognition rate of the new method for 17 types of radar signals with different modulation modes can reach 97.1%, which is about 2.4 percentage points higher than that of the dilated residual network; when the signal-to-noise ratio is −10 dB, the recognition rate can reach 92.1%, which is about 0.7 percentage points higher than that of EfficientNet, and improves the recognition rate of complex multi-type radar signal modulation modes under low signal-to-noise ratio conditions.
You Might Like
Recommended ContentMore
Open source project More
Popular Components
Searched by Users
Just Take a LookMore
Trending Downloads
Trending ArticlesMore