Information-Theoretic Pilot Design for Downlink Channel Estimation in FDD Massive MIMO Systems

• Published in: IEEE transactions on signal processing Vol. 67; no. 9; pp. 2334 - 2346
• Main Authors: Yujie Gu, Zhang, Yimin D.
• Format: Journal Article
• Language: English
• Published: New York IEEE 01.05.2019; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Antennas; Base stations; Channel estimation; Computer simulation; Covariance matrices; Covariance matrix; Downlink; frequency division duplex (FDD); Frequency division duplexing; Gaussian distribution; Gaussian mixture distribution; Grassmannian manifold; Information theory; information-theoretic metric; Manifolds (mathematics); massive multiple-input multiple-output (MIMO); MIMO (control systems); MIMO communication; pilot design; Symbols; Training; Wireless communications
• ISSN: 1053-587X; 1941-0476
• DOI: 10.1109/TSP.2019.2904018

Massive multiple-input multiple-output (MIMO) is one of the most promising techniques for next generation wireless communications due to its superior capability to provide high spectrum and energy efficiency. Considering the very large number of antennas employed at the base station, however, the pilot overhead for downlink channel estimation becomes unaffordable in frequency division duplex (FDD) multiuser massive MIMO systems. In this paper, we propose an information-theoretic metric to design the pilot for downlink channel estimation in FDD multiuser massive MIMO systems. By exploiting the low-rank nature of the channel covariance matrix, we first derive the minimum number of pilot symbols required to ensure perfect channel recovery, which is much less than the number of antennas at the base station. Further, under a general channel model that the channel vector of each user follows a Gaussian mixture distribution, the pilot symbols are designed by maximizing the weighted sum of the Shannon mutual information between the measurements of the users and their corresponding channel vectors on the complex Grassmannian manifold. Simulation results demonstrate the effectiveness of the proposed information-theoretic pilot design for the downlink channel estimation in FDD massive MIMO systems.