Game-Theoretic Framework for Cooperative Sensing and Fair Spectrum Access in Multichannel Cognitive Radio Networks.
Yuan Lu, Alexandra Duel‐Hallen · arXiv (Cornell University) · 2016
In cognitive radio (CR) networks, secondary users (SUs) sense the spectrum to identify and possibly transmit over temporarily unoccupied channels licensed to primary users (PUs). When the received PU signal is weak, spectrum sensing by individual SUs becomes unreliable. To improve the sens- ing accuracy, SUs can cooperate to discover idle time slots. However, fair payoff (spectrum access) allocation has not been investigated for cooperative sensing in multichannel CR networks with heterogeneous PU signal quality at different SU sensors although such scenarios are prevalent in practice. In this paper, a two-layer coalitional game model is investigated, in which the coalition formation and the payoff allocation problems are decoupled: in the top-layer hedonic game, SUs form disjoint coalitions over potentially available channels while in the bottom- layer game, SUs negotiate over the payoff allocation within each coalition. The cooperatively detected spectrum opportunities are then shared among the coalition members in a coordinated man- ner according to their binding agreement on payoff allocation. The proposed game is decentralized and fosters cooperation by providing each SU with the transmission opportunities it deserves. Moreover, we propose a new physical-layer approach to distribute the network-level miss-detection (MD) constraints fairly among the interfering SUs for guaranteed PU protection. We demonstrate the performance advantages of the AND-rule combining of spectrum sensing results for heterogeneous SUs. Numerical results show that the proposed game outperforms previously investigated collaborative sensing and multichannel access approaches in terms of energy efficiency, throughput, SU fairness, and computational complexity. Finally, the effect of mobility on the computational complexity of the proposed game is discussed.