Near optimal linear precoder for multiuser MIMO for discrete alphabets

In this paper we look at the effect of discrete constellation alphabets on linear precoding for the downlink of multiuser (MU) MIMO in the context of LTE. We underline the fundamental difference in the approach of precoding if the alphabets are assumed to be discrete constellations rather than the idealized Gaussian assumption. We show that the problem of finding global optimal linear precoder taking into account discrete inputs is non-convex and we propose a method of finding a near optimal linear precoder. Underlining the viability of MU MIMO for future wireless communications as LTE, we further propose in this paper a precoding strategy based on low resolution LTE precoders which necessitate 2 bits feedback from the users. The proposed strategy encompasses geometrical interference alignment at eNodeB and the use of low complexity MU detectors at the users. On one hand, this strategy relegates the interference seen by each user by a geometric scheduling algorithm while on the other hand, users exploit the structure of this interference in the detection process. Simulation results validate improved performance of the proposed strategy over single user schemes.