Multiuser space-time channel estimation for CDMA under reduced-rank constraint

This paper addresses the problem of estimating the multiuser channels in CDMA systems when using a linear array of antennas. The training sequences to be used for multiuser channel estimation have limited lengths. In order to control the errors of the (unstructured) estimate of space-time (S-T) channel we propose to reduce the complexity when parameterizing the channel. This is achieved by constraining each matrix that models the S-T channel of each user to have a low-rank. The rank of the matrix accounts for the degrees of space and/or time diversity in the S-T channel. Here we propose to estimate the reduced-rank channel simultaneously for all the users so as to control the multiaccess interference (i.e., all the users are considered signals for each other). The covariance matrix of the intercell interference is constrained accordingly. Simulation results for realistic propagation/interference environments confirm the expected benefits.