A novel spectrum selection strategy for matching multi-service secondary traffic to heterogeneous primary spectrum opportunities

In order to increase spectrum utilization efficiency, CRs (Cognitive Radios) have been introduced to reuse white spaces left unused by legacy services under the strict constraint of not interfering them. In this context, this paper proposes to exploit a statistical characterisation of Primary User (PU) activity to be retained in Radio Environment Maps (REMs) for spectrum selection purposes. The objective is to match multi-service secondary traffic to heterogeneous primary spectrum opportunities minimizing the SpHO (Spectrum handOver) rate. Specifically focusing on dependence structures potentially exhibited by primary ON/OFF periods, two spectrum selection criteria have been first proposed to benchmark the utility of the embedded statistical patterns in the REM. Results have shown that the one or the other criterion can introduce significant gains with respect to a random selection depending on the secondary configuration and characteristics of PUs. Therefore, a novel pro-active spectrum selection strategy combining the proposed criteria has been developed and proven to achieve in most of the cases the best performance for a given secondary service mix and the dependence level between primary ON/OFF periods.