DFT-based Channel Estimation Techniques for Massive MIMO Systems

Abstract

In massive MIMO systems, efficient and highly accurate channel state information (CSI) at the base station are essential requirements for tackling the effect of pilot contamination to achieve the potential benefits of the systems. In this paper, we propose two discrete Fourier transform (DFT)-based channel estimation techniques for massive MIMO systems. The proposed methods mitigate the pilot contamination significantly via modifying the DFT-based estimation through iterations and most significant (MST) approaches. The simulation results obtained through these approaches demonstrated the effectiveness of the DFT-based channel estimation techniques in alleviate/eliminate the pilot contamination when compared to conventional channel estimators, in terms of channel estimation accuracy and achievable uplink rate. Index Terms-Massive MIMO; channel estimation; discrete Fourier transform (DFT).

High Performance Interference Suppression in Multi-User Massive MIMO Detector

“…Finally, MMSE-OSIC with the optimized user sorting and the SR-K-best demonstrates performance close to the maximum likelihood algorithm in the additive white noise channel, i.e. in case of i = 0 in (10). However, the mentioned MMSE-OSIC is quite sensitive to external interference from unknown users (when i = 0), while the basic MMSE algorithm is robust to the correlated noise due to the IRC algorithm with the matrix R uu .…”

“…where w is the weight vector, h is the channel estimation vector as described in [10] and [11], y = y 1 y 2 y 3 T , covariance matrix R yy of the received signal is defined similarly to (2). The matrix R yy can also be calculated as:…”

Data-Aided LS Channel Estimation in Massive MIMO Turbo-Receiver

“…Therefore, CE plays a key role in Massive MIMO efficiency. Finally, there is about 1dB...2dB theoretically proven performance loss compared to an ideal channel estimation, which is much more than CE losses of 0.1dB...0.5dB in 4G technology as shown in [4], [5] and [6].…”

“…We utilize (144,288) LDPC code with the Min-Sum decoding algorithm in the receiver end [17], [18] . The DFT-based channel estimation was implemented as described in [6]. QuaDRiGa, short for "QUAsi Deterministic RadIo channel GenerAtor" [16], was used to generate realistic radio channel responses in system-level simulations of 5G scenarios.…”

Theoretical Performance Bound of Uplink Channel Estimation Accuracy in Massive MIMO

“…The number of antennas in Massive MIMO starts from 64 while in 4G with a common MIMO this number is limited by 8 [2]. With a growing number of antennas correlation between them gets higher, which provides extra abilities to enhance performance significantly via joint processing as shown in [3,4,5], but practical performance is still too far from the performance, achieved with ideal CE. As a result, CE topic attracts many researches to compensate gap between ideal and practical CE performances [6].…”