[Paper Review] Asymptotic Analysis of Large Intelligent Surface Assisted MIMO Communication

Qurrat-Ul-Ain Nadeem, Abla Kammoun+3 more|arXiv (Cornell University)|Mar 19, 2019

Advanced Wireless Communication TechnologiesEngineering24 references50 citations

TL;DR

This paper proposes a large intelligent surface (LIS)-assisted MIMO system that enhances downlink spectral efficiency by passively reflecting signals via $N$ phase-shifting elements, achieving massive MIMO-like performance with far fewer active antennas. For rank-one and full-rank LoS channels between base station and LIS, it derives asymptotic optimal linear precoding and phase optimization, showing SINR gains even as $K$ grows, with simulations confirming significant energy reduction and performance comparable to massive MIMO.

ABSTRACT

This work focuses on the downlink of a single-cell multi-user (MU) system in which a base station (BS) equipped with $M$ antennas communicates with $K$ single-antenna users through a large intelligent surface (LIS) installed in the line-of-sight (LoS) of the BS. LIS is envisioned to offer unprecedented massive multiple-input multiple-output (MIMO) like gains by utilizing $N$ passive reflecting elements that induce phase shifts on the impinging electromagnetic waves to smartly reconfigure the signal propagation environment. We study the minimum signal-to-interference-plus-noise ratio (SINR) achieved by the optimal linear precoder (OLP), that maximizes the minimum SINR subject to a given power constraint for any given LIS phase matrix, for the cases where the LoS channel matrix between the BS and the LIS is of rank-one and of full-rank. In the former scenario, the minimum SINR is bounded by a quantity that goes to zero with $K$. For the high-rank scenario, we develop accurate deterministic approximations for the parameters of the asymptotically OLP, which are then utilized to optimize the LIS phase matrix. Simulation results show that the LIS-assisted system can offer massive MIMO like gains with a much fewer number of active antennas, thereby significantly reducing the energy consumption at the BS.

Motivation & Objective

• To analyze the asymptotic performance of a large intelligent surface (LIS)-assisted multi-user MIMO system in a single-cell downlink scenario.
• To investigate the minimum signal-to-interference-plus-noise ratio (SINR) achieved by the optimal linear precoder (OLP) under power constraints for different LIS phase matrix configurations.
• To derive deterministic approximations for the asymptotic OLP parameters in the high-rank LoS channel case to enable efficient LIS phase matrix optimization.
• To demonstrate that LIS can deliver massive MIMO-like spectral efficiency gains while drastically reducing base station energy consumption.
• To quantify the impact of LIS channel rank (rank-one vs. full-rank) on system performance and SINR scaling with user count $K$.

Proposed method

• Models the downlink of a single-cell multi-user MIMO system with a base station (BS) using $M$ active antennas and $K$ single-armed users, assisted by a large intelligent surface (LIS) with $N$ passive reflecting elements.
• Assumes line-of-sight (LoS) propagation between BS and LIS, with the LoS channel matrix between them either rank-one or full-rank.
• Analyzes the minimum SINR achieved by the optimal linear precoder (OLP) under a total transmit power constraint for a given LIS phase matrix.
• Derives deterministic equivalents for the asymptotic OLP parameters in the full-rank case using random matrix theory, enabling practical phase optimization.
• Proposes a phase matrix optimization framework based on the derived deterministic approximations to maximize the minimum SINR.
• Validates the theoretical findings via simulations, comparing performance against conventional massive MIMO in terms of spectral efficiency and energy efficiency.

Experimental results

Research questions

• RQ1 How does the minimum SINR scale with the number of users $K$ when the LoS channel between the BS and LIS is rank-one?
• RQ2 What are the asymptotic characteristics of the optimal linear precoder (OLP) in the full-rank LoS channel scenario between the BS and LIS?
• RQ3 Can accurate deterministic approximations be derived for the OLP parameters in the high-SNR and large-$N$ regime to enable efficient LIS phase optimization?
• RQ4 To what extent can an LIS-assisted system achieve massive MIMO-like spectral efficiency gains with significantly fewer active antennas at the BS?
• RQ5 How does the performance of the LIS-assisted system compare to traditional massive MIMO in terms of energy efficiency and SINR scaling with $K$?

Key findings

• In the rank-one LoS channel case, the minimum SINR achieved by the optimal linear precoder (OLP) is bounded by a quantity that tends to zero as the number of users $K$ increases.
• For the full-rank LoS channel case, the paper derives accurate deterministic approximations for the asymptotic OLP parameters, enabling effective phase matrix optimization.
• The proposed LIS-assisted system achieves massive MIMO-like spectral efficiency gains even with a much smaller number of active antennas at the base station.
• Simulation results confirm that the LIS-assisted system significantly reduces energy consumption at the BS while maintaining high spectral efficiency.
• The minimum SINR is maximized through optimized LIS phase shifts, with performance gains becoming more pronounced as $K$ increases.
• The theoretical analysis and simulation results demonstrate that LIS can reconfigure the propagation environment to achieve performance comparable to massive MIMO with lower hardware cost and power consumption.