Pilot Design for Gaussian Mixture Channel Estimation in Massive MIMO

Abstract

Massive multiple-input multiple-output (MIMO) is a promising technique for 5G communications due to its superior spectrum and energy efficiencies. Despite its many advantages, the high number of antennas used in massive MIMO brings many challenges in practical implementations. Among them, the pilot overhead for downlink channel estimation becomes unaffordable in frequency division duplex (FDD) massive MIMO systems. In this paper, we exploit the available a priori knowledge of the channel to optimize the pilot design. By utilizing the low-rank nature of the channel matrix, we first derive the minimum number of pilot symbols required for perfect channel recovery. Further, under the general Gaussian mixture model for the channel vector, the pilot symbols are optimized to maximize the mutual information between the measurements of the user and the corresponding channel vector. Simulation results demonstrate the effectiveness of the proposed optimal pilot design for the downlink channel estimation in FDD massive MIMO systems.