LS-Based Pilot Contamination Pre-coding for Large-scale MIMO System

Abstract

To reduce pilot contamination, a downlink precoding algorithm was put forward for Large-scale MIMO system. It set the Least Square (LS) criterion to the objective function, and then solved the precoding matrices with the Lagrange multiplier method. In the solving, it utilized not only the estimated channel information but also the pilot contamination information, which was acquired from some blank resource elements, so could more precisely reduce pilot contamination. Based on simulation results, it was shown that the proposed algorithm can achieve better performance than popular Large-scale MIMO precoding algorithms, e.g. single-cell Zero-Forcing precoding and multi-cell MMSE precoding. It could break through the limit of pilot contamination and acquire approximately linear performance improvement with the number of antennas increasing.