Distributed Learning-Based Multi-Band Multi-User Cooperative Sensing in Cognitive Radio Networks
Multi-band cooperative spectrum sensing can provide access to a wide range of spectrum in cognitive radio networks (CRNs). The design of multi-band spectrum sensing is very challenging mainly due to scheduling of secondary users (SUs) to sense a subset of channels. In this paper, we propose a distributed learning-based multi-band multi-user cooperative spectrum sensing (M2CSS) scheme to select most appropriate SUs to sense channels. The proposed scheme allows SUs to sense multiple channels, and consists of two stages: 1) leader selection for each channel, and 2) selection of corresponding cooperative SUs to sense these channels. We formulate an optimization problem to select leaders that can effectively communicate with other SUs subject to the constraint that each SU can act as a leader for only one channel, and there will be only one leader for each channel. We then formulate another optimization problem to select corresponding cooperative SUs for each channel. After this stage, selected cooperative SUs sense channels, and use consensus learning to determine the availability of channels in a distributed manner. Simulation results show that the proposed M2CSS scheme can enhance detection performance, avoid the choice of redundant cooperative SUs, owning similar sensed information, and provide fair energy consumption for all channels compared to the existing schemes.
|, , ,|
|2018 IEEE Global Communications Conference, GLOBECOM 2018|
|Organisation||Department of Systems and Computer Engineering|
Gharib, A. (Anastassia), Ejaz, W. (Waleed), & Ibnkahla, M. (2019). Distributed Learning-Based Multi-Band Multi-User Cooperative Sensing in Cognitive Radio Networks. In 2018 IEEE Global Communications Conference, GLOBECOM 2018 - Proceedings. doi:10.1109/GLOCOM.2018.8648118