Modular Precoding Based on Zero Forcing in Large Scale Antenna Systems with Imperfect Channel State Information

In Large scale antenna systems (LSAS), the user number is usually much less than the number of base station antennas. By leveraging this characteristic, Modular precoding based on zero forcing (MZF) is proposed to achieve higher spectrum efficiency than Maximum ratio combining (MRC), while preserve the advantage of distributive processing. The achievable rates of the MZF, MRC and Zero forcing (ZF) are derived under the generalized model of imperfect Channel state information (CSI). The imperfect CSI model includes the channel estimation delay, the noise in the pilot channel, and the pilot contamination. The analytical achievable rates match the simulations even with a small number of base station antennas. When the antenna number approaches to infinity, the analytical achievable rates of MZF converge to the analytical achievable rates of ZF.