Eavesdropping on DS-CDMA signal

This paper considers the problem of signal eavesdropping in direct-sequence code-division multiple-access (DS-CDMA) channels. We aim at extracting the information data symbols of all concurrent users in a DS-CDMA system without the knowledge of spreading sequences (signature waveforms), and channel state information (CSI) as well as no availability of pilot signaling (training sequence). We first develop an iterative least-squares (ILS) core procedure to blindly extract the information symbols of the concurrent users. Then, enhanced ILS (E-ILS), a procedure described herein in detail that is based on statistical analysis of repeated independent ILS processing of observation signal, is seen to offer most effective signal eavesdropping. Simulation studies show that the proposed E-ILS signal eavesdropping algorithm can achieve probability of error close to what may be attained with known spreading sequences and CSI.