Performance of adaptive multiuser detectors in space-time coded CDMA systems

In this paper, we present a comparative performance evaluation of adaptive multiuser detectors, including stochastic gradient (SG) and recursive least squares (RLS) algorithms (which require training data), and minimum output energy (MOE) and subspace-based MMSE (S-MMSE) algorithms (which do not require training data), under near-far conditions in a space-time coded CDMA system. We show that, in a near-far multituser scenario, increasing the number of transmit antennas degrade the near-far resistance performance of the SG and MOE detectors significantly at high near-far ratios, to the extent that the diversity benefit of multiple transmit antennas is nullified. The RLS and S-MMSE algorithms, on the other hand, are shown to maintain their near-far resistance without loosing much on the diversity gain, when the number of transmit antennas is increased.