Cognitive Radios in Cooperative Environment: Detection, Sensing and Clustering of Spectral Bands

The cognitive radio allows spectrum sharing between licensed users (primary users) and many unlicensed users (secondary users). The secondary users are only allowed to use a licensed spectrum in the absence of its rightful owner. Thus, the ability to accurately sense the presence of the rightful owners is highly essential. In this paper, a novel cooperative spectrum sensing (CSS) approach to efficiently sense a wideband spectrum is introduced. This paper addresses two specific issues in CSS. The first issue is on dynamic detection of primary user´s bands. A dynamic band clustering (DBC) algorithm that uses K-means clustering technique is proposed in this paper. The proposed algorithm reduces the number of erroneous narrow sub-bands resulting from spurious noise. This in turn minimizes the number of sub-bands to be detected and hence the overall sensing time. The second issue is on combining the local decisions made by different secondary users. A novel dynamic weighting decision fusion algorithm is also proposed in this paper. Based on extensive simulation studies, the proposed fusion algorithm is shown to be superior to conventional decision fusion techniques.