An Iterative Semi-Blind Channel Estimation Scheme and Uplink Spectral Efficiency of Pilot Contaminated One-Bit Massive MIMO Systems

Massive multiple-input multiple-output (MIMO) systems consist of a large number of antennas and radio frequency (RF) chains, which cause enormous circuit power consumption at the RF front ends. Since high-resolution analog-to-digital converters are the main contributor in the RF circuit power consumption, one-bit massive MIMO is considered as one of the potential solutions. The channel state information (CSI) acquisition in the presence of pilot contamination is a challenging task in such systems. In this paper, we present an iterative semi-blind-based channel estimation scheme for pilot contaminated one-bit multi-cell multi-user massive MIMO systems, which comprise of two stages: initialization and iteration. The initial channel estimate uses pilots in the initialization stage, which is further refined in the iteration stage with the help of both pilot and a few data symbols. We derive lower bounds on the uplink achievable rate with both perfect and imperfect CSI availability at the base station. Through simulations, we show that the proposed estimator achieves a considerable improvement in mean square error, bit-error-rate, power efficiency, and spectral efficiency from one-bit pilot-aided estimators at the cost of a nominal increase in computational complexity. Moreover, the proposed scheme in one-bit system achieves almost the same spectral efficiency as that of a pilot-aided estimator in an infinite-resolution-based massive MIMO system for lower values of signal-to-noise-ratio (SNR).

I. Introduction

Massive multiple-input multiple-output (MIMO) is an integral part of the fifth generation communication systems [1], [2]. It is a large antenna system with an equivalent number of radio frequency (RF) chains, which increases the hardware complexity and circuit power consumption [3]. The difficulty in practical implementability of massive MIMO is its huge circuit power consumption. The high-resolution analog-to-digital converters (ADCs) of RF chains are one of the major contributors to power consumption [4]. Thus, low resolution (1–3 bit) ADCs are seen as one of the potential solutions in reducing the power requirement of massive MIMO systems [5], [6]. In addition, systems with one-bit ADCs do not require automatic gain control and linear amplifiers, which further reduce RF circuit cost and hardware complexity [7]. This work considers massive MIMO using one-bit ADCs in their RF chains which are commonly known as one-bit massive MIMO. Recently, the usage of low-bit quantizers in massive MIMO has received significant attention [8]–[10]. The uplink (UL) spectral efficiency (SE) of massive MIMO systems with low-resolution ADCs is investigated in [8]–[10]. The authors of [8], [11] show that the large array gain of massive MIMO compensates the capacity loss incurred with the usage of one-bit quantizer at high SNRs. The literature [8]–[10] assume perfect channel state information (CSI) for capacity analysis of low-resolution based massive MIMO systems, which is an impractical and unrealistic scenario. In practice, CSI acquisition is a necessary and important prerequisite for the practical realization of one-bit massive MIMO systems.