Distributed Partial Collaborative Deep Learning for Spectrum Sensing

With the rapid development of wireless communications, the number of mobile devices is increasing at an unprecedent speed, leading to a shortage of spectrum resources. Dynamic spectrum allocation (DSA) is an effective way to alleviate the scarcity of spectrum resources. DSA relies on spectrum sensing, which aims to detect unoccupied frequency bands. Traditional spectrum sensing methods only consider the scenario of a single sensing node, which can only monitor a limited geographical scope. In order to monitor a large geographical range, it is necessary to consider the distributed sensing architecture. Due to the varying hardware accuracy and sensing environment, there exists a heterogeneity among devices in a distributed system. To address this issue, this paper proposes a new distributed spectrum sensing architecture. The proposed architecture consists of several sensing nodes, each of which is equipped with a convolutional neural network (CNN) in order to identify whether the monitored spectrum is occupied. The shallow layers of these CNNs are demanded to be the same, while the deep layer of each CNN is independently trained using each node's local training data. The proposed method can significantly enhance the sample efficiency while enabling each CNN to be well-adapted to the local SNR (signal-to-noise ratio). Simulation results demonstrates the efficiency of the proposed method.