Chi-square QQ plot technique for detecting multivariate anomalies in centralized wideband cooperative compressive sensing in the presence of malicious users

Abstract

This paper presents a novel centralized cooperative scheme which combines compressed sensing (CS) method based on the analog-to-information converter (AIC) with statistical and graphical chi-square Quantile-Quantile (QQ) plot detector. The received signal at every cognitive radio (CR) user in the presence of Additive-White-Gaussian-Noise (AWGN) and intentional interference attack signal (jamming) is transformed in to a digital signal using an Analog-to-Information Converter (AIC). This new scheme consists for detecting anomalies problem using a new method of detection so-called chi-square QQ plot by compressed observations of every cognitive radio (CR) receiver collected in sampling matrix. Cooperative via CR sensing receivers enables the CRs to detect attacks in the presence of AWGN and jamming attack signals. Furthermore, as an interest key point, collaboration between all CR users makes match possible to sample more compressively at every CR receiver. Two hypotheses H 0 and H 1 are proposed in this paper to know if there is problem of anomalies in the spectrum. Finally, the results obtained show better performance.