Fast Discovery of Spectrum Opportunities in Cognitive Radio Networks

Abstract

We address the problem of rapidly discovering spectrum opportunities for seamless service provisioning for secondary users (SUs) in cognitive radio networks (CRNs). Specifically, we propose an efficient sensing-sequence that incurs a small opportunity-discovery delay by considering (1) the probability that a spectrum band (or a channel) may be available at the time of sensing, (2) the duration of sensing on a channel, and (3) the channel capacity. We derive the optimal sensing-sequence for channels with homogeneous capacities, and a suboptimal sequence for channels with heterogeneous capacities for which the problem of finding the optimal sensing-sequence is shown to be NP-hard. To support the proposed sensing-sequence, we also propose a channel-management strategy that optimally selects and updates the list of backup channels. A hybrid of maximum likelihood (ML) and Bayesian inference is also introduced for flexible estimation of ON/OFF channel-usage patterns and prediction of channel availability when sensing produces infrequent samples. The proposed schemes are evaluated via in-depth simulation. For the scenarios we considered, the proposed suboptimal sequence is shown to achieve close-to-optimal performance, reducing the opportunity-discovery delay by up to 47% over an existing probability-based sequence. The hybrid estimation strategy is also shown to outperform the ML-only strategy by reducing the overall opportunity-discovery delay by up to 34%.