Novel channel estimation methods under pilot contamination in Massive MIMO

Abstract

Massive multiple-input multiple-output (Massive MIMO) is a key feature of proposed 5G cellular systems, offering potentially many benefits. However, all benefits rest on the ability to obtain channel state information (CSI) at the base station (BS) during uplink transmission. The BS can in principle measure CSI from known user-transmitted pilot sequences, but in a multiple cell system, the use of non-orthogonal pilot sequences in different cells leads to a problem of pilot contamination. In this thesis, we mainly focus on recently proposed covariance-aided channel estimation for time division duplex (TDD) Massive MIMO cellular systems suffring pilot contamination. The recent work assumes that the covariance matrices of users are known and do not change with time. In this thesis, we address two new scenarios: 1) the covariance matrix in one cell changes due to a new user arriving in that cell; 2) the covariance matrices of all users change due to mobility of the users. In these scenarios, we develop novel algorithms that estimate the new covariance matrices and then use these estimated covariance matrices to obtain high quality channel estimates.The proposed algorithms are compared with other estimation methods mentioned in this thesis to show the benefits of the proposed algorithms.