Achieving cooperative spectrum sensing in wireless cognitive radio networks

  • Authors:
  • Chengqi Song;Qian Zhang

  • Affiliations:
  • Hong Kong University of Science and Technology;Hong Kong University of Science and Technology

  • Venue:
  • ACM SIGMOBILE Mobile Computing and Communications Review
  • Year:
  • 2009

Quantified Score

Hi-index 0.00

Visualization

Abstract

Dynamic spectrum access has been studied to exploit instantaneous spectrum availability by opening licensed spectrum to secondary users. To achieve high spectrum efficiency, secondary unlicensed users need to continuously sense spectrum to detect the presence of primary licensed users. Cooperative spectrum sensing has been recognized as a powerful solution to improve spectrum sensing performance, which requires nearby wireless nodes to share sensing results with each other. However, information sharing is achieved through broadcasting in wireless networks, which can provide free-riding opportunity for selfish nodes. Selfish nodes can benefit from receiving the sensing results from its neighbors by free without sharing. Therefore, appropriate strategies are essential to enforce and sustain the cooperation among neighboring nodes. In this paper we model cooperative spectrum sensing as an N-player horizontal infinite game, and study varies strategies for it. In wireless networks, the frequently occurred collisions make the cooperation enforcement problem quite challenging as it is hard to tell whether the information lost is due to nodes' selfishness or wireless collision. In this paper, we prove that Grim Trigger strategy, a classical strategy to stimulate cooperation in an infinite game, can result in poor performance due to random errors. We then propose a strategy basing on Carrot-and-Stick strategy, which can recover cooperation among multiple players from deviation. We prove that if nodes are sufficiently far-sight, or equivalently the entire system runs sufficiently long, the Nash Equilibrium of the proposed strategy for spectrum sensing game is still mutual cooperation, even under collision situation. We also prove that the proposed strategy is robust to collisions and colluding cheat.