Robust Energy Detection Based on Bayesian Estimation for Cognitive Radio

The energy detection is recently studied extensively as a promising candidate for primary signals detection in cognitive radio. However, one fundamental disadvantage of energy detection is that its performance degrades in the presence of inaccurate knowledge of noise power (noise uncertainty) which inevitably occurs in practical implementations. In this paper, we introduce a novel Bayesian estimation based energy detection (BEED) which, confirmed by simulations, can almost completely eliminate detrimental effects of noise uncertainty. In addition, the consistency of BEED is proved which means that it will correctly detect primary users with probability one when the sampling number increases to infinity.