[Paper Review] Large Antenna Analysis of Multi-Cell Full-Duplex Networks

Jingwen Bai, Ashutosh Sabharwal|arXiv (Cornell University)|Jun 16, 2016

Full-Duplex Wireless Communications27 references3 citations

TL;DR

This paper analyzes uplink and downlink ergodic rates in multi-cell full-duplex MIMO networks with large-antenna base stations, accounting for practical impairments like pilot contamination, channel estimation error, and imperfect self-interference cancellation. It demonstrates that the 2× spectral efficiency gain of full-duplex over half-duplex is preserved in the large-antenna regime, enabling significant spectral efficiency with fewer antennas in finite systems.

ABSTRACT

We study a multi-cell multi-user MIMO full-duplex network, where each base station (BS) has multiple antennas with full-duplex capability supporting single-antenna users with either full-duplex or half-duplex radios. We characterize the up- and downlink ergodic achievable rates for the case of linear precoders and receivers. The rate analysis includes practical constraints such as imperfect self- interference cancellation, channel estimation error, training overhead and pilot contamination. We show that the 2X gain of full-duplex over half-duplex system remains in the asymptotic regime where the number of BS antennas grows infinitely large. We numerically evaluate the finite SNR and antenna performance, which reveals that full-duplex networks can use significantly fewer antennas to achieve spectral efficiency gain over the half-duplex counterparts. In addition, the overall full-duplex gains can be achieved under realistic 3GPP multi-cell network settings despite the increased interference introduced in the full-duplex networks.

Motivation & Objective

• To characterize uplink and downlink ergodic achievable rates in multi-cell full-duplex MIMO networks with large-antenna base stations.
• To account for practical impairments such as pilot contamination, channel estimation error, training overhead, and imperfect self-interference cancellation.
• To evaluate whether the 2× spectral efficiency gain of full-duplex over half-duplex is preserved in the asymptotic regime as the number of base station antennas grows large.
• To assess finite-SNR and finite-antenna performance, showing that full-duplex networks can achieve gains with significantly fewer antennas than half-duplex counterparts.

Proposed method

• Derives uplink and downlink ergodic achievable rate expressions using linear precoders and receivers under imperfect channel state information.
• Models interference from full-duplex operation, including inter-cell and intra-cell interference from uplink users and neighboring base stations.
• Applies large-system analysis techniques, leveraging central limit theorem approximations and Wishart matrix properties to compute expectations of channel gain products.
• Uses MMSE channel estimation to model estimation error and derives variance expressions for effective channel gains under imperfect CSI.
• Incorporates training overhead and pilot contamination by modeling pilot contamination as a zero-mean complex Gaussian random variable with known variance.
• Derives closed-form expressions for achievable rates by computing expectations over channel estimates and error terms, using results from random matrix theory.

Experimental results

Research questions

• RQ1 Does the 2× spectral efficiency gain of full-duplex over half-duplex persist in a multi-cell massive MIMO setting with practical impairments?
• RQ2 How do pilot contamination and channel estimation error affect the ergodic rate in full-duplex multi-cell networks with large-antenna base stations?
• RQ3 Can the full-duplex gain be maintained in the asymptotic regime as the number of base station antennas tends to infinity?
• RQ4 What is the finite-SNR and finite-antenna performance trade-off in full-duplex networks compared to half-duplex systems?
• RQ5 Can the overall full-duplex gain be achieved under realistic 3GPP multi-cell network conditions despite increased interference?

Key findings

• The 2× spectral efficiency gain of full-duplex over half-duplex is asymptotically preserved when the number of base station antennas grows large, even under imperfect self-interference cancellation.
• In the asymptotic regime, the impact of intra-cell and inter-cell interference, as well as channel estimation error, vanishes as the number of antennas increases.
• The 2× gain is maintained when only full-duplex users are served, but not when half-duplex users are present due to training overhead and pilot contamination effects.
• Finite-SNR simulations show that full-duplex networks can achieve the same spectral efficiency as half-duplex systems with significantly fewer antennas.
• Despite increased interference, the overall full-duplex gain is achievable in realistic 3GPP multi-cell network scenarios, validating practical feasibility.