Wireless Internet. 6th International ICST Conference, WICON 2011, Xi’an, China, October 19-21, 2011, Revised Selected Papers

Research Article

Demand-Matching Spectrum Sharing in Cognitive Radio Networks: A Classified Game

Download
377 downloads
  • @INPROCEEDINGS{10.1007/978-3-642-30493-4_51,
        author={Shaohang Cui and Jun Cai},
        title={Demand-Matching Spectrum Sharing in Cognitive Radio Networks: A Classified Game},
        proceedings={Wireless Internet. 6th International ICST Conference, WICON 2011, Xi’an, China, October 19-21, 2011, Revised Selected Papers},
        proceedings_a={WICON},
        year={2012},
        month={10},
        keywords={Cognitive radio distributed spectrum sharing classified game correlated equilibrium no-regret learning},
        doi={10.1007/978-3-642-30493-4_51}
    }
    
  • Shaohang Cui
    Jun Cai
    Year: 2012
    Demand-Matching Spectrum Sharing in Cognitive Radio Networks: A Classified Game
    WICON
    Springer
    DOI: 10.1007/978-3-642-30493-4_51
Shaohang Cui1,*, Jun Cai1,*
  • 1: University of Manitoba
*Contact email: shaohang@ee.umanitoba.ca, jcai@ee.umanitoba.ca

Abstract

Cognitive Radio (CR) has been proposed as a promising technique to solve spectrum scarcity problem in wireless communications. For the implementation of CR, one major challenge is to design distributed spectrum sharing, which needs to efficiently coordinate CRs in accessing the spectrum opportunistically based on only local information. To address this problem, in this paper, we make use of the heterogeneity among users in cognitive radio networks (CRNs) and propose a distributed cooperative game with classified players. A prioritized CSMA/CA technique is adopted so that CRs select channels and their priority to access channel based on their satisfaction history, a public signal for CRs to collaborate to achieve the Correlated Equilibrium (C.E.). A no-regret learning algorithm is adopted to learn the C.E. Simulation results show that the proposed C.E. based classified game (CECG) can achieve up to 40% better performance compared to the unclassified one.