Efficient In-Band Spectrum Sensing Using Swarm Intelligence for Cognitive Radio Network

Spectrum sensing mechanisms enable cognitive radio networks to detect primary users (PUs) and utilize spectrum holes for secondary user (SU) transmission. However, precise PU detection leads to longer sensing time and lower achievable throughput. In this paper, we propose a particle swarm optimization (PSO)-based scheme for an in-band local spectrum sensing to address the tradeoff between sensing time and throughput. Using methodological analysis, a fast convergence PSO (FC-PSO) scheme is derived by implementing a distribution-based stopping criterion subject to detection performance, optimization time, and SU gain. At the target probability of detection of at least 90%, the results show significant improvements of ∼ 45% for sensing time, 70% for the probability of false alarm, and 12% for achievable throughput compared with nonoptimal sensing scheme at signal-to-noise ratio of 0 dB. FC-PSO also outperforms other optimization schemes in terms of convergence speed. The proposed scheme is proved to be an energy-efficient solution for practical implementation as it outperforms the other algorithms in terms of lower computational complexity as well as providing the best tradeoff values in meeting the objective function of sufficient opportunistic access for an SU under optimized sensing time for maximized throughput, while providing high protection to the PU.