Space-time minimum variance receivers for CDMA systems

Abstract

The use of an antenna array in code division multiple access (CDMA) base station can effectively improve system capacity, quality, and coverage of wireless communication systems. This makes space-time signal processing for CDMA systems a promising technology. In general, space-time based CDMA receivers use knowledge about the received signature waveform across space and time for the desired user. The channel information is normally acquired through the use of training sequences (or symbols) known both by the transmitter and receiver. In this dissertation, we propose several space-time linear receivers based on certain optimality criteria: space-time matched filter (MF) receiver, spacetime minimum mean squared error (MMSE) receiver, and space-time minimum variance (MV) receiver for an antenna array CDMA system. The space-time MV receiver is based on a certain constrained optimization criterion which does not require the knowledge of the direction of any user, and does not require the spreading code of the interfering users. For this constrained optimization problem, a new constraint matrix which is composed of the spreading code of the desired user is introduced, and a batch algorithm is proposed for determining the unknown parameter vector and the weight vector of the receiver. Algorithms for recursively updating the optimum solutions of the space-time MMSE receiver and the space-time MV receiver based on the recursive least squares (RLS) adaptation method are developed. The robust space-time MV receiver is derived in the presence of modeling errors or mismatch due to unmodeled channel effects. The resulting proposed receiver is based on the signal and noise subspace decomposition of the signal covariance matrix, and is insensitive to the mismatch. Simulation results are presented for demonstrating the improved bit error rate (BER) performance of the proposed space-time receivers in a multipath channel. The receivers also show superior performance to the conve...