4th International ICST Conference on Wireless Internet

Research Article

Low-overhead Scheduling Algorithms for OFDMA Relay Networks

Download452 downloads
  • @INPROCEEDINGS{10.4108/ICST.WICON2008.4938,
        author={Karthikeyan Sundaresan and Xiaodong Wang and Mohammad Madihian},
        title={Low-overhead Scheduling Algorithms for OFDMA Relay Networks},
        proceedings={4th International ICST Conference on Wireless Internet},
        publisher={ICST},
        proceedings_a={WICON},
        year={2010},
        month={5},
        keywords={Algorithms relays diversity spatial reuse scheduling performance},
        doi={10.4108/ICST.WICON2008.4938}
    }
    
  • Karthikeyan Sundaresan
    Xiaodong Wang
    Mohammad Madihian
    Year: 2010
    Low-overhead Scheduling Algorithms for OFDMA Relay Networks
    WICON
    ICST
    DOI: 10.4108/ICST.WICON2008.4938
Karthikeyan Sundaresan1, Xiaodong Wang1, Mohammad Madihian1
  • 1: Broadband and Mobile Networking, NEC Labs America

Abstract

Enhanced data rates and connectivity are key requirements for providing ubiquitous mobile access in next-generation cellular networks. Relay-enabled cellular networks, marked by their adoption in IEEE 802.16j standard, have become a viable candidate in such an endeavor. Such relay networks not only provide multi-user and (OFDM) channel diversity gains that are available in conventional cellular systems, but also provide spatial reuse gains, arising from the simultaneous transmissions on different hops of the network on the same channel. However, the efficient exploitation of these gains, calls for intelligent scheduler design at the BS that must not only accommodate the multi-hop nature of the network, but also address the resulting significant overhead incurred in the form of feedback. In this work, we present relay-assisted scheduling algorithms that efficiently exploit the available diversity and spatial reuse gains at the cost of minimal feedback overhead. The proposed solutions improve performance over conventional approaches by over 50% along with a scalable feedback overhead that grows only with the number of relays in the network and not with the number of users.