Optimization of Dynamic Double-threshold Cooperative Spectrum Sensing Performance in Cognitive Radio Networks

As a key technology of cognitive radio, spectrum sensing has been extensively researched. The main parameters which measure the performance of cooperative spectrum sensing are the global false alarm probability and the global undetected probability, and the sum of them is defined as global error probability. In this paper, the optimization of cooperative spectrum sensing based on double threshold energy detection is studied. First of all, the voting threshold of the voting fusion rule is optimized under the circumstance of fixed detection thresholds, which makes the global error probability reach the minimum value. Then the detection thresholds of energy detection are optimized under the optimal voting fusion rule. The optimal detection thresholds are dynamic under different SNR conditions, so the optimal detection thresholds are determined by the SNR, under different SNR, the global error probability always reached the minimum value, so the performance of cooperative spectrum sensing gets improved. Simulation results show that the lowest global error probability is achieved when the voting threshold of the voting fusion rule and the detection thresholds of energy detection reach their optimum values respectively.