Robust Collaborative Spectrum Sensing over non-ideal reporting channels

Abstract

Collaborative Spectrum Sensing (CSS) is well known for its applicability to combat fading and shadowing effects in Cognitive Radio (CR) networks. In this paper, we show the ineffectiveness of conventional CSS techniques over non-ideal reporting channels when no prior knowledge about the primary user's signal exists. Then, we propose a two-sample distributionfree test based on rank summation of decisions. The proposed method, in spite of conventional methods, does not require any apriori statistical information about the reporting channel's fading and/or noise model. The results of the numerical study indicated that the proposed method outperforms conventional equal-gain combining (EGC) CSS technique over non-ideal reporting channel when there is no prior know.