A Blind Distributed Spectrum Sensing Scheme With Homogeneity Test

This paper proposes a new scheme for blind distributed spectrum sensing (DSS), where multiple distributed cognitive radios (CRs) and a fusion center (FC) collaboratively detect the availability of a frequency band of interest under communications constraints without knowledge on the channels, signal power, or noise power. By following the energy detection and homogeneity test concepts, in the proposed scheme, the CRs make local binary decisions based on relative comparison of the total energy in a sampling window, particularly, whether a CR has the maximum total energy. With independent channels, the use of the maximum function generates desirable discrepancy among the CR decisions when the frequency band is in use. The introduction of the window size enables the balancing between communications costs and the performance. Closed-form results on the distribution of the CR decisions are obtained for Gaussian noises and signals. Further, asymptotic analytical expressions are derived on the detection performance of the proposed scheme for both the generalized likelihood ratio test (GLRT) and $\chi ^{2}$ -test at the FC. Simulation results are shown which validate the theoretical results and show the advantage of the proposed scheme to others.