Fast Cooperative Spectrum Sensing Based on Sequential Probability Ratio Test

Abstract

Cooperation can improve the performance of spectrum sensing in the cognitive radio networks by taking advantage of diversity gains. It has been shown that sequential probability ratio test (SPRT) is a promising data fusion technique for cooperative spectrum sensing. However, most exiting works assume that the primary spectrum status stays unchanged during the cooperative sensing process. We argue that, in reality, the spectrum state can vary at anytime during one round of cooperative sensing, and this kind of spectrum variation could cause severe distortion to the performance of SPRT. In this paper,we propose a novel cooperative sensing scheme to cop with this problem. In the proposed scheme, we introduce a multi-process mechanism into the traditional SPRT technique, which can effectively deal with the variation of spectrum within a sensing process. Besides, we also study the design tradeoff issue in our scheme. Simulation results show that our multi-process SPRT scheme can improve the response speed to spectrum status variation and show a better performance in cooperative spectrum sensing.