The application discloses a
spread spectrum communication interference sensing method based on an attention mechanism, which comprises the following steps: 1, collecting common interference
signal types in a
spread spectrum communication task and modeling; 2, modeling a channel environment, constructing a received
signal model according to a
spread spectrum communication
system, generating a large number of interference-free signals and interfered signals under different jamming
signal-to-signal ratios, and constructing a training
data set; 3, establishing an interference sensing
Transformer network, which is composed of an embedding module, a twin
transformer module and a gating module, wherein the twin
transformer module comprises two transformers, which are respectively used for learning the correlation characteristics between subcarriers and the correlation characteristics between real parts and imaginary parts; 4, training the interference sensing
Transformer network constructed by using the training
data set to obtain network parameters; and 5, fixing the network parameters, inputting a real-time spread spectrum received signal spectrum, and completing interference type identification by the network.