Impact of noise estimation on eigenvalue based spectrum sensing in cognitive radio systems

In order to gain awareness of available transmission opportunities, a cognitive radio [1] system requires the implementation of a spectrum sensing unit, which must indicate whether a transmission is taking place in the considered channel. In order to provide such indication, a binary hypothesis testing experiment must be periodically performed [2]. Many algorithms have been proposed for the computation of such test statistics [3]. In this paper we focus on the performance of Hybrid Approach of semi-blind eigenvalue-based detection (EBD) algorithms, especially RLRT [4], in which the noise variance is estimated in a given number of auxiliary slots.