[Paper Review] Experimental Performance Evaluation of Cell-free Massive MIMO Systems Using COTS RRU with OTA Reciprocity Calibration and Phase Synchronization

Yang Cao, Pan Wang+10 more|arXiv (Cornell University)|Aug 30, 2022

Advanced MIMO Systems Optimization4 citations

TL;DR

This paper proposes an over-the-air (OTA) reciprocity calibration framework for cell-free massive MIMO using commercial off-the-shelf (COTS) RRUs, enabling accurate downlink channel state information (CSI) estimation via frequency-domain reference signals and a genetic algorithm-based RRU grouping. The key contribution is a prototype system demonstrating robust performance under calibration delay and phase drift, with local precoding showing insensitivity to delay when data streams are limited.

ABSTRACT

Downlink coherent multiuser transmission is an essential technique for cell-free massive multiple-input multiple output (MIMO) systems, and the availability of channel state information (CSI) at the transmitter is a basic requirement. To avoid CSI feedback in a time-division duplex system, the uplink channel parameters should be calibrated to obtain the downlink CSI due to the radio frequency circuit mismatch of the transceiver. In this paper, a design of a reference signal for over-the-air reciprocity calibration is proposed. The frequency domain generated reference signals can make full use of the flexible frame structure of the fifth generation (5G) new radio, which can be completely transparent to commercial off-the-shelf (COTS) remote radio units (RRUs) and commercial user equipments. To further obtain the calibration of multiple RRUs, an interleaved RRU grouping with a genetic algorithm is proposed, and an averaged Argos calibration algorithm is also presented. We develop a cell-free massive MIMO prototype system with COTS RRUs, demonstrate the statistical characteristics of the calibration error and the effectiveness of the calibration algorithm, and evaluate the impact of the calibration delay on the different cooperative transmission schemes.

Motivation & Objective

• To address the challenge of downlink CSI acquisition in cell-free massive MIMO systems without CSI feedback.
• To overcome RF hardware mismatch-induced reciprocity violation in distributed RRU deployments.
• To develop a transparent, COTS-compatible calibration method compatible with 5G NR frame structure.
• To evaluate the impact of calibration delay and phase drift on joint and local precoding schemes in a real-world prototype.

Proposed method

• Design of a frequency-domain reference signal for OTA reciprocity calibration, fully transparent to COTS UEs and RRUs.
• Introduction of an averaged Argos calibration algorithm for group-based RRU calibration with low complexity.
• Employment of interleaved RRU grouping with a genetic algorithm to optimize calibration performance across distributed RRUs.
• Implementation of a cell-free massive MIMO testbed using 5G COTS RRUs and real-time channel estimation.
• Adoption of both joint precoding (JP-RZF) and local precoding (L-RZF) for performance comparison under calibration delay.
• Statistical analysis of calibration error and phase drift effects using measured data from the prototype system.

Figure 1: Configuration of a special slot for OTA reciprocity calibration. D, U and G denote the states in the OFDM symbol for the transmitting, receiving and guard periods, respectively.

Experimental results

Research questions

• RQ1 How can OTA reciprocity calibration be implemented transparently in COTS-based cell-free massive MIMO systems without modifying commercial devices?
• RQ2 What is the impact of calibration delay on the spectral efficiency of joint and local precoding schemes in distributed CF-mMIMO?
• RQ3 How does phase drift between RRUs affect system performance, particularly under local precoding?
• RQ4 Can a group-based calibration approach with genetic algorithm optimization achieve sufficient accuracy for practical deployment?
• RQ5 Under what conditions is local precoding robust to calibration delay and hardware mismatch?

Key findings

• The prototype system achieved stable calibration with a phase drift range of (-π/6, π/6), resulting in only 0.3 dB SNR loss and ~0.1 bps/Hz SE degradation at high SNR.
• Local precoding (L-RZF) was found to be robust to calibration delay when the number of data streams was less than or equal to the number of antennas per RRU.
• Joint precoding (JP-RZF) was highly sensitive to calibration delay, especially when the number of data streams approached the maximum spatial degrees of freedom.
• With a 20-ms calibration delay, the performance gap between JP-RZF and L-RZF was minimal, indicating that L-RZF maintains stable performance under practical delay conditions.
• The EDU-based L-RZF outperformed JP-RZF under high calibration delay, suggesting that localized processing can mitigate delay-induced performance loss.
• The averaged Argos calibration algorithm achieved a favorable trade-off between complexity and accuracy in multi-RRU calibration.

Figure 3: Two-symbol CARS.