New Detection Methods of DS-CDMA Signal at Lower SNR

Abstract

DS-CDMA in the modern civil and military communications is widely used. Due to its low power spectral density, the DS-CDMA signal always floods in the background noise, bringing difficulties to the signal’s reconnaissance. Therefore, DS-CDMA signal detection has attracted attention of researchers. In this paper, the fourth-order moment chip (m4x(0,0,t)) of DS-CDMA signals is derived, against the background of Gaussian white noise, and get the conclusion that (m4x(0,0,t) can suppress Gaussian white noise.  In order to reduce the estimation error of (m4x(0,0,t), the quadratic fourth-order moment chip is proposed and finally two detection methods based on the quadratic fourth-order moment chip of DS-CDMA are proposed. Experimental results show that the detection performance of the methods is better than that of the double frequency method, spectrum-reprocessing method and cepstrum method. When the false alarm probability is 1%, the detection probability of presented methods reaches above 90% at SNR -22dB.