Spectrum Sensing Algorithm Based on LSTM and Its Implementation of Multiple USRP
Huachao Lu, Zhijin Zhao · 2019
Aiming at the problem that the fusion rules of cooperative spectrum sensing have great impact on performance, a cooperative sensing algorithm based on LSTM, which is implemented on multiple USRPs is proposed. The received signal has different sequence characteristics when the primary user signal is present or absent. LSTM is used to extract the temporal characteristics of each primary user's signal sequence, and the fully connected layer is used to fuse the features in the fusion center, then softmax is used to classify fusion features. A number of USRPs and a host are built a spectrum sensing system, and the LSTM model obtained by offline training is used to perform online real-time detection. The system can effectively detect the primary user signal.