Random Matrix Theory-Based Analysis and Design of Semi-Blind Processing for Massive MIMO Systems

Overview

Massive multiple-input multiple-output (MIMO) systems have emerged as a key enabling technology for future wireless communication networks due to their ability to significantly improve spectral efficiency, energy efficiency, and communication reliability. The performance of massive MIMO systems, however, critically depends on the availability of accurate channel state information (CSI). Conventional pilot-based channel estimation methods often require excessive pilot overhead, particularly in systems with large antenna arrays. To alleviate this issue, semi-blind channel estimation has attracted considerable attention as an effective approach that jointly exploits pilot symbols and received data observations to improve channel estimation accuracy while reducing pilot overhead. Nevertheless, the analysis and design of semi-blind channel estimation methods remain challenging in massive MIMO systems, especially in large-dimensional regimes where conventional finite-dimensional approaches become analytically intractable.

The asymptotic behaviors of deterministic and stochastic Cramér–Rao bounds (CRBs) for semi-blind channel estimation are first analyzed under various asymptotic regimes, and tractable closed-form expressions are derived to characterize the tradeoff between pilot overhead and channel estimation accuracy. CRB-based resource allocation strategies for multi-user uplink systems are further investigated under both joint and sequential estimation frameworks. In addition, a robust semi-blind channel estimation method based on a regularized least-squares formulation is developed by combining pilot-based information with blind subspace criteria. Furthermore, the impact of imperfect CSI on uplink receiver performance is investigated through asymptotic signal-to-interference-plus-noise ratio (SINR) analysis for linear receivers under both pilot-based and semi-blind channel estimation schemes. Closed-form asymptotic approximations are derived and validated through numerical simulations under realistic non-terrestrial network (NTN) scenarios, demonstrating that semi-blind channel estimation can significantly improve receiver performance and reduce pilot overhead compared with conventional pilot-based approaches.