Multi-stage cross entropy optimization algorithm for hard combining schemes in cognitive radio network

Spectrum sensing optimization is the process of finding the optimal set of sensing parameters in order to maximize the optimization objective while meet the restrictions imposed. The detection accuracy of a cognitive radio network (CRN) improves through using a cooperative spectrum sensing (CSS) scheme. However, increasing the number of SU necessitates a growth in the cooperation overhead of the system leading to degradation the throughput of the CRN. Multi stage-cross entropy (MSCE) optimization algorithm has been proposed to optimize the trade-off between global probability of detection at fusion center (FC) and achievable throughput in cooperative CRNs, and then compared the results with genetic algorithm (GA) and particle swarm optimization (PSO) algorithms. The proposed approach is based on cross entropy (CE) optimization method. A bi-objective (BO) function have been formulated for static PU signal state scenarios. Numerical results show that the MSCE performance is superior in terms of achievable PU detection rate when compared with GA, PSO and hard decision combining (HDC) rules. Additionally, the BO-MSCE optimization system based-HDC rules achieve a best fitness score higher than that of the GA and PSO for the OR, AND, and Majority rules, respectively.