Performance Analysis of a Semiblind Channel Estimator in One-Bit Massive MIMO Systems

Massive multiple-input multiple-output (MIMO) suffers from RF hardware complexity, cost, and circuit power consumption. The high-resolution analog-to-digital converter (ADC) at the RF front end is one of the major contributors in power depletion. Thus, one-bit massive MIMO becomes an effective solution. However, the availability of accurate channel state information (CSI) at the base station remains a challenging task. In this article, we propose a semiblind channel estimator for one-bit massive MIMO systems and evaluate its uplink (UL) performance. The proposed algorithm consists of two stages: initialization and iteration. The initial channel estimate is obtained from a pilot-aided initialization stage, which is refined in the iteration stage with the help of both pilot and a few data symbols. The semiblind approach simultaneously improves the estimation accuracy and spectral efficiency with minimum pilot length. We also derive a theoretical lower bound on the UL achievable rate under perfect and imperfect CSI. Through simulations, we show that the semiblind scheme achieves a considerable improvement in mean square error, bit-error-rate, spectral efficiency, and bit energy against the existing one-bit pilot based estimators at the cost of a nominal increase in computational complexity.