Novel distributed sequential nonparametric tests for spectrum sensing

Abstract

We consider the problem of spectrum sensing in cognitive radio networks under model uncertainties, i.e., the distribution under null hypothesis and the alternate hypothesis is not known. We model the presence of the primary signal as a mean change problem. Two simple nonparametric sequential algorithms for testing the mean are proposed. These nonparametric algorithms are applied to a decentralized detection framework. These algorithms show better performance than some of the existing nonparametric tests in literature. The performance of these algorithms is analysed via theory and simulations.

Distributed nonparametric sequential spectrum sensing under electromagnetic interference

“…Asymptotic analysis of the random walk test is provided in [37]. In the next section we briefly present that and also include the effects of heavy tailed noise and fading which was not discussed in [37].…”

“…Asymptotic analysis of the random walk test is provided in [37]. In the next section we briefly present that and also include the effects of heavy tailed noise and fading which was not discussed in [37]. This will explain why Mrandom walk and M 2 -random walk perform better under heavy-tailed EMI and outliers and using truncation on H improves performance in the presence of slow fading.…”