Joint interference suppression and multiuser detection for three-cell multiple relay communications

Abstract

To cope with interference in three-cell multiple relay systems, this paper proposes schemes that detect two users' signals while linearly suppressing one remaining user's signal. The proposed schemes do not require channel state information (CSI) feedback and can drastically reduce the computational complexity compared to the maximum likelihood detection (MLD), which is the optimal signal detection and detects all three users' signals. As one proposed scheme, a minimum mean-squared error (MMSE) scheme estimates weight coefficients for the interference suppression under the MMSE criterion, and thus needs to estimate CSI of all the users by using all the three preamble sequences like MLD. Another proposed scheme called a recursive least squares (RLS) scheme directly estimates the weight coefficients by the RLS algorithm. In contrast with the MMSE scheme and MLD, the RLS scheme needs to know only two preamble sequences and thus is more feasible. Computer simulations of orthogonal frequency-division multiplexing (OFDM) transmission under three-cell and frequency selective fading conditions show that the proposed schemes can maintain good bit error rate in contrast with a conventional scheme for two-cell systems, and that the RLS scheme can outperform the MMSE scheme with estimated CSI when decision-directed estimation with error selection is applied.