Blind Channel Estimation and Multi-User Detection for Wireless CDMA Systems

In this paper, we propose a novel system architecture for channel estimation and multi-user detection in CDMA systems. Estimation and detection are performed jointly by sequential basis selection algorithms; namely the basic matching pursuit (BMP) algorithm and the orthogonal matching pursuit (OMP) algorithm. We start by demonstrating that the iterations of the BMP algorithm are equivalent to that of the well known detection method, successive interference cancelation (SIC) in code domain. We then introduce the OMP algorithm as a multi-user detector. This novel detection structure orthogonalizes the SIC process hence improves the detection performance. By using simulation results, we demonstrate that its performance is equivalent to decorrelating detector. Inspired from these results, we propose a joint blind channel estimation and detection scheme employing OMP. This novel scheme is less complex than subspace based techniques and gives better performance.