Fast Discovery of Spectrum Opportunities in Cognitive Radio Networks

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