02-05-2023 дата публикации
Номер: CN116056092A
Принадлежит:
The invention belongs to the technical field of wireless communication, and discloses a cooperative spectrum sensing method based on federated learning, which comprises three parts of data preprocessing, model training and updating and online detection. According to the method, historical received data is processed, a corresponding training set is constructed to train a neural network, and model parameters are continuously updated by using a federated learning algorithm, so that consumption of communication resources is reduced, a model can quickly extract a hidden rule of data, and a reasonable joint detection value calculation method is designed; according to the method, the false alarm probability is taken into consideration to solve a proper detection threshold, sensing results of signals received by different nodes in real time are fused by using a neural network, the error rate of the sensing results of a single channel is reduced, and real-time judgment of a spectrum occupancy state ...
Подробнее