On Optimal Threshold Selection in Cooperative Spectrum Sensing for Cognitive Radio Networks: An Energy Detection Approach Using Fuzzy Entropy Maximization

Spectrum sensing with improved reliability is the key for realizing dynamic spectrum access through cognitive radio systems. In this paper, an optimization problem is formulated to determine multiple threshold values for cooperative spectrum sensing in cognitive radio networks. Fuzzy entropy maximization on sensed energy is suggested as solution. Two methods, namely selection combining (SC) and maximum ratio combining (MRC) at fusion center are studied to obtain the value of the resultant sensing energy. A set of parameters on energy values are used to define different fuzzy functions that indicate different degree of presence and absence of primary user (PU). The parameters are then searched for optimal values that maximize fuzzy entropy using differential evaluation algorithm. Simulation results at low signal-to-noise ratio (SNR) are shown to highlight the efficacy of using multiple threshold values over single threshold. Simulation results also demonstrate the effects of increase in receiver diversity and the number of PU samples on detection reliability. Results shown shed a light on the choice of SC or MRC depending on the value of operating SNR to resolve the trade-off in detection reliability and energy requirement.