A sparse signal perspective for blind user identification in multiuser DS-CDMA

We present a blind user identification method for the direct-sequence code division multiple access (DS-CDMA) signal from a sparse signal representation perspective using an overcomplete basis composed of all possible spreading sequences. The main idea of the proposed algorithm is to combine the sparsity characteristic of the DS-CDMA signal in terms of spreading sequences with the autocorrelation and the cross-correlation characteristics of spreading sequences. Firstly, the DS-CDMA signal samples are cut into several temporal windows. Each signal of the temporal window is corresponding to a measurement of the active user's spreading sequence. Secondly, the overcomplete dictionary of spreading sequences is constructed according to the structure of spreading sequence and the support coefficient of its every atom is calculated by performing the Orthogonal Matching Pursuit (OMP) method on samples of all temporal windows. Finally, with the help of the minimum description length (MDL) criterion, the number of active users is estimated and then the spreading sequences of active users are identified by inspecting their support coefficients. Simulation results demonstrate that our proposed method is effective for both synchronous short-code (SC) DS-CDMA signal and long-code (LC) DS-CDMA signal, even with many active users and low SNR.