Linear Precoding for MIMO With LDPC Coding and Reduced Complexity

In this paper, the problem of designing a linear precoder for Multiple-Input Multiple-Output (MIMO) systems employing Low-Density Parity-Check (LDPC) codes is addressed under the constraint of minimizing the dependence between the system´s receiving branches, thus reducing the relevant transmitter and receiver complexities. Our approach constitutes an interesting generalization of Bit-Interleaved Coded Modulation with Multiple Beamforming (BICMB) which has shown many benefits in MIMO systems. We start with a Pareto optimal surface modeling of the system and show the difficulty involved in the corresponding optimization problem. We then propose an alternative, practical technique, called Per-Group Precoding (PGP), which groups together multiple input symbol streams and corresponding receiving branches in the “virtual” channel domain (after singular value decomposition of the original MIMO channel), and thus results in independent transmitting/receiving streams between groups. We show with numerical results that PGP offers almost optimal performance, albeit with significant reduction both in the precoder optimization and LDPC EXIT chart based decoding complexities.