Spectrum sensing based on goodness of fit test with unilateral alternative hypothesis

Abstract

Several classical goodness of fit (GoF) tests have been used in spectrum sensing. A new GoF test with a unilateral alternative hypothesis for spectrum sensing is proposed. The new test statistic is derived and extended to the case of finite samples. It is shown that the proposed spectrum-sensing method provides improvement in performance and reduction in computational complexity.

Unilateral left‐tail Anderson Darling test‐based spectrum sensing with Laplacian noise

“…Based on the GoF theory, several fitting criteria have been applied in spectrum sensing, such as the Cramervon Mises (CM) test, the Kolmogorov–Smirnov (KS) test, the order statistic test, the Anderson Darling (AD) test etc [21–31]. Nevertheless, the proposed detectors in [21–29] have assumed Gaussian noise. Thus it may not be appropriate to employ them directly in the presence of Laplacian noise.…”

Unilateral right‐tail Anderson‐Darling test based spectrum sensing for cognitive radio

“…Recently, goodness-of-fit (GoF) test has gained much attention due to its effectiveness in spectrum sensing when the prior information of the primary signal and the fading channel is completely or partly unavailable. Representative GoF tests for spectrumsensing are Kolmogorov-Smirnov, Cramer-von Mises, Anderson-Darling (AD) and unilateral AD (UAD) test (see [1,2] and references therein). In [1], an AD test based spectrum sensing (AD scheme) is proposed where the spectrum sensing is formulated as a bilateral chi-square distribution GoF test problem.…”

“…In [2], spectrum sensing is formulated as the following unilateral chi-square distribution GoF test problem given the fact that x i under H 1 containing primary signals and noises simultaneously is larger than x i under H 0 containing only noises, namely…”

“…The proof is completed. □ Simulations: In this section, the proposed URAD scheme is compared with AD [1] and UAD [2]. Fig.…”

Contribuições ao Sensoriamento Espectral em Canais Cognitivos Sujeitos a Desvanecimento

“…O teste de hipóteses também pode ser formulado em função da Função Cumulativa de Probabilidade (FCP) do sinal observado: se a FCP F n (x) das amostras obtidas pelo rádio cognitivo (também referenciada como FCP empírica) adere ou não à função cumulativa do sinal de referência F 0 (x) [104] • H 0 , se F n (x) = F 0 (x);…”