Cross entropy algorithm for improved soft fusion-based cooperative spectrum sensing in cognitive radio networks

In cooperative spectrum sensing of cognitive radio based network, various methods of soft decision fusion (SDF) and hard decision fusion (HDF) schemes have been proposed to optimize the performance of detecting primary users so that they are well-protected from harmful cognitive radio access. In this paper, cross entropy (CE) based algorithm is proposed as an efficient technique for optimizing the weighting coefficients vector of an SDF-based cooperative spectrum sensing scheme. The proposed CE based algorithm is compared with existing deterministic methods as well as with an evolutionary-based genetic algorithm (GA) method. Simulation results show that the proposed CE scheme outperforms the other schemes in terms of the achievable fitness of primary users' detection probability, convergence, and stability.