Exploiting Multi-Channel Diversity in Spectrum-Agile Networks

Recently, spectrum-agile (SA) networks have been recognized as a viable solution to the spectrum-scarcity problem in spectrum-scarcity problem. In SA networks, secondary (unlicensed) users are allowed to opportunistically utilize idle licensed spectrum bands, thus improving spectrum utilization efficiency and accommodating more users and applications. We take a two- step approach to the problem of maximizing the throughput of an SA network. The first step is to determine a subset of "candidate" channels that a secondary device will consider for its channel-switching. The candidate channels are selected based on their estimated utilization. We then propose channel-aware switching to determine when and where to switch to, among the candidate channels. Wireless channels are assumed to experience independent Rayleigh fading, and modeled with a finite-state Markov channel (FSMC) model. Our evaluation results show that the proposed channel-aware switching scheme significantly outperforms the traditional forced-switching scheme in terms of average throughput.