A NOVEL WEIGHTED SNR BASED COOPERATIVE SPECTRUM SENSING SCHEME IN COGNITIVE RADIO NETWORKS

Abstract

Cognitive radio solves the problem of sparsely available radio spectrum by introducing optimum utilization of spectrum. The rst step in designing the cognitive radio is identi- cation of spectrum hole. This is carried out using di erent spectrum sensing techniques. Spectrum sensing techniques can be broadly divided into hard and soft sensing. Hard sensing schemes include OR, AND, Majority, general N out of K rule and soft sensing schemes include square law selection, selection combining, equal gain combining. State of art technique multiselective cooperative sensing had been proposed in a recent paper. The optimization of multiselective scheme had also been discussed. This optimization has very high computational complexity. Moreover, looking at the ROC of multiselective scheme and the upper bound further improvement of ROC can be carried out. To overcome these limitations a new soft sensing scheme called weighted SNR scheme has been proposed. The analytical expressions for detection and false alarm probability under i.i.d. Rayleigh fading have been derived. Simulations have been performed in MATLAB to compare the performance of proposed scheme with multiselective scheme and other existing schemes. The ROC of proposed scheme is superior to multiselective scheme as well as all the conventional techniques. For a particular case, the total error rate of proposed scheme is found out to be 0.034 as compares to 0.077 for multiselective scheme. Simulations have been performed to verify the results. Further sensitivity analysis of the proposed scheme has been performed to study the e ect of di erent parameters on the performance of this scheme. It has been found that after a particular value of time- bandwidth product the ROC saturates. The proposed scheme is thus found to be a better alternative to other soft sensing schemes under i.i.d. Rayleigh fading channel.