Pilot contamination precoding for interference reduction in large scale antenna systems

A large scale antenna system (LSAS) entails a large number of base station (BS) antennas serving a much smaller number of terminals. LSAS demonstrates large gain in spectral-efficiency and energy-efficiency compared with conventional MIMO schemes. In the regime of infinite number of BS antennas, the performance of LSAS can be improved unboundedly by using a multi-cell cooperation method called pilot contamination precoding (PCP). As the number of BS antennas grows PCP completely cancels inter-cell interference resulted from unavoidable use of nonorthogonal pilot sequences in neighboring cells. In this paper, we consider downlink PCP design in the finite antenna regime, where besides pilot contamination, three other types of interference become prominent. We formulate the PCP design as maximizing the minimum signal to interference-plus-noise ratio (SINR) subject to network sum power constraint. We propose optimal and simple suboptimal algorithms for finding PCP matrices. We show that in a practical scenario, say with M = 100 antennas, these approaches give very significant improvement over conventional LSASs.