PaperlayerView exact revision on arXiv

Source-linked AI summary

Overview of Full-Dimension MIMO in LTE-Advanced Pro

Hyoungju Ji, Younsun Kim, Juho Lee, Eko Onggosanusi, Younghan Nam, Jianzhong Zhang, Byungju Lee, Byonghyo Shim

arXiv:1601.00019v4cs.IT

TL;DR

FD-MIMO addresses the demand for higher cellular spectral and energy efficiency using large antenna arrays, but practical deployment involves hardware, CSI, and pilot challenges. This article surveys FD-MIMO features and Release 13 standardization, emphasizing realistic design issues and evaluations. It reports field-trial potential and a 105% cell-edge gain for type I over conventional 8Tx LTE-A in a cited comparison.

  • Problem

    Future cellular systems seek higher spectral and energy efficiency, while massive-MIMO commercialization faces hardware, deployment, fronthaul, and high-dimensional CSI challenges.

  • Method

    The article provides an overview of FD-MIMO, covering Release 13 standardization, system features, realistic CSI and channel-design issues, and typical scenario evaluations.

  • Results

    105% cell-edge gain is reported for type I relative to conventional 8Tx LTE-A in the cited FD-MIMO comparison.

  • Takeaways & Limitations

    FD-MIMO’s practical integration into LTE requires attention to CSI acquisition, pilot overhead, two-dimensional active arrays, and realistic three-dimensional propagation.

Abstract

from arXiv · show

Multiple-input multiple-output (MIMO) systems with a large number of basestation antennas, often called massive MIMO, have received much attention in academia and industry as a means to improve the spectral efficiency, energy efficiency, and processing complexity of next generation cellular system. Mobile communication industry has initiated a feasibility study of massive MIMO systems to meet the increasing demand of future wireless systems. Field trials of the proof-of-concept systems have demonstrated the potential gain of the Full-Dimension MIMO (FD-MIMO), an official name for the MIMO enhancement in 3rd generation partnership project (3GPP). 3GPP initiated standardization activity for the seamless integration of this technology into current 4G LTE systems. In this article, we provide an overview of the FD-MIMO system, with emphasis on the discussion and debate conducted on the standardization process of Release 13. We present key features for FD-MIMO systems, a summary of the major issues for the standardization and practical system design, and performance evaluations for typical FD-MIMO scenarios.

I. INTRODUCTION

FD-MIMO extends LTE toward larger two-dimensional antenna arrays, while addressing practical CSI, pilot-overhead, hardware, and deployment constraints. The article reviews Release 13 standardization, system features, realistic channel conditions, and performance scenarios.

  • System scope: FD-MIMO is 3GPP’s name for the LTE MIMO enhancement targeting up to 64 transmitter antenna ports.3GPP began standardization to integrate the technology into LTE-Advanced Pro.
  • Article focus: The article surveys Release 13 standardization, including TXRU architectures, beamformed CSI-RS, 3D beamforming, CSI feedback, and realistic FD-MIMO evaluations.It emphasizes practical issues and the standardization discussion rather than only ideal pilot and feedback conditions.
  • CSI acquisition: CSI feedback overhead increases with the number of transmit antennas, while CSI distortion degrades multiuser-MIMO precoding performance.The feedback amount must scale with N_T to control quantization error.
  • CSI acquisition: With N_T = 64, orthogonal reference-signal overhead occupies approximately 48% of downlink resources.The resulting pilot burden reduces resources available for data transmission.
  • 2D active antenna system: A two-dimensional active antenna array enables beam control in both elevation and azimuth while accommodating many antennas in limited deployment space.A dual-polarized 8 × 8 array is described as requiring 1.0 × 0.5 m.
  • 3D channel environment: Realistic FD-MIMO channel models must represent three-dimensional propagation, including height- and distance-dependent LOS, pathloss, and elevation spread.The paper notes less pathloss on higher floors and changing elevation spread with user height and distance.

D. RS transmission for CSI acquisition

FD-MIMO replaces conventional CSI acquisition approaches with beamformed CSI-RS, reducing feedback and pilot overhead as antenna counts grow while improving reference-signal quality.

  • Reference-signal roles: DM-RS cannot be used for CSI measurements because it is present only on scheduled time/frequency resources.DM-RS is UE-specific and precoded with the same weight as data transmission.
  • Beamformed CSI-RS: Beamformed CSI-RS is introduced as a new FD-MIMO reference signal for CSI acquisition.It supports channel training through transmitter architectures that control phase and gain in digital and analog domains.
  • Feedback overhead: Feedback for beamformed CSI-RS scales logarithmically with the number of reference signals, unlike channel-vector quantization feedback, which is proportional to N_T.The resulting feedback advantage is pronounced when N_T is large.
  • Pilot overhead: Non-precoded CSI-RS pilot overhead increases with N_T and can substantially reduce FD-MIMO sum capacity, whereas beamformed CSI-RS overhead is proportional to N_B and independent of N_T.Consequently, its rate loss remains marginal as N_T increases.
  • Reference-signal quality: 4.3dB gain in signal power is provided by beamformed CSI-RS over non-precoded CSI-RS when N_T = 32 and N_B = 12.The comparison assumes the same transmit power is distributed across fewer beamformed reference signals.
  • Feedback mechanism: Beamformed CSI-RS requires a new feedback mechanism because conventional codebooks cannot measure beamformed transmissions.The eNB therefore needs feedback designed for the beamformed transmission process.

III. SYSTEM DESIGN AND STANDARDIZATION OF FD-MIMO SYSTEMS

Release 13 targets FD-MIMO with larger two-dimensional antenna arrays and coordinates standardization around antenna hardware, channel measurement, feedback, and multi-user support.

  • Standardization targets: Release 13 initially targets systems supporting up to 16 antennas, while support for more than 16 antennas is deferred to subsequent releases.The broader study identifies issues for systems with up to 64 transmit antennas in a two-dimensional array.
  • Standardization targets: The study discusses 2D arrays, elaborated TXRUs, enhanced channel measurement and feedback schemes, and up to eight co-scheduled users.These features define the main design dimensions considered for FD-MIMO standardization.
  • CSI standardization: CSI measurement and feedback are tightly coupled to the FD-MIMO standardization effort.The paper treats this coupling as a central issue alongside the other system-design topics.

A. Deployment scenarios

FD-MIMO evaluation considers 3D urban macro and micro deployments, where antenna height and propagation conditions determine useful vertical beam steering and user scheduling strategies.

  • Deployment scenarios: The paper defines 3D-UMa with rooftop antennas and 3D-UMi with antennas below the rooftop.These are the two typical deployment scenarios used for FD-MIMO design and evaluation.
  • 3D-UMa: Down-tilted transmission is desirable in 3D-UMa because rooftop diffraction dominates propagation.Different steering angles can separate channels associated with multiple UEs.
  • 3D-UMi: In 3D-UMi, direct signal paths are dominant because users are located above the antenna height, enabling up- and down-tilting to schedule UEs on different floors.The scenario typically has a smaller cell radius than 3D-UMa.
  • 3D-UMi: More UEs can be co-scheduled in 3D-UMi without increasing inter-user interference because line-of-sight conditions predominate.The paper also notes that line-of-sight probability increases in 3D-UMa when height differences between the eNB and UEs are considered.

B. Antenna configurations

FD-MIMO uses element-level two-dimensional active arrays whose vertical and horizontal dimensions, polarization, and spacing shape beam patterns, interference control, compatibility, and deployment constraints.

  • Active antenna arrays: Active antennas use low-power amplifiers at each element and require element-level modeling because radiation patterns depend on element count and spacing.The array is characterized by vertical and horizontal element counts and polarization degree.
  • Array structure: The benchmark array uses dual polarization with P = 2, M = 8 vertically at 0.8λ spacing, and N = 4 horizontally at 0.5λ spacing.M, N, and P denote vertical elements, horizontal elements, and polarization degree, respectively.
  • Beam patterns: Vertical scheduling is more effective for controlling inter-user interference because the elevation null direction is 11° versus 30° horizontally.The smaller vertical null direction provides finer separation in the elevation domain.
  • Array structure: A tall or fat array structure is favorable because it generates a sharp beam, although environmental changes can reduce flexibility.The preferred geometry depends on the intended beam pattern and operating conditions.
  • Backward compatibility: The suggested configuration preserves backward compatibility by using the same total antenna-element count as conventional 8Tx systems.The vertical configuration maintains cell coverage, while the horizontal configuration supports conventional LTE MIMO operation.
  • Deployment constraints: Large antenna spacing can increase inter-cell interference for cell-edge UEs through narrow beamforming, an effect called the flash-light effect.Design parameters should therefore account for user location, cell radius, building height, and antenna height.

C. TXRU architectures

TXRU architectures connect radio units to antenna elements through either partitioned or connected array structures, supporting conventional or beamformed CSI-RS transmission. These architectures determine how CSI-RS measurements and precoding weights are formed.

  • Architecture components: TXRU architecture comprises the TXRU array, antenna array, and radio distribution network that delivers transmit and received signals.The radio distribution network connects power amplifiers to antenna elements and antenna elements to low-noise amplifiers.
  • CSI-RS transmission: Array partitioning is associated with conventional CSI-RS transmission, whereas array connected architecture is associated with beamformed CSI-RS transmission.The two architectures differ in how CSI-RS resources are assigned and how the UE obtains channel information.
  • Array partitioning: Array partitioning divides antenna elements into groups, connecting each TXRU to one group and assigning orthogonal CSI-RS resources.The UE measures each TXRU’s channel from its CSI-RS transmission.
  • Array connected architecture: Array connected architecture links multiple TXRUs to each antenna element through additional RF combining circuitry.This structure supports mixing RF signals from multiple TXRUs before transmission.
  • Beamformed CSI-RS: With beamformed CSI-RS, the UE measures the precoded channel hv, whose SNR is maximized at the target direction.The beamformed observation is y = hvx+n.

D. New CSI-RS transmission strategy

Release 13 considered conventional non-precoded and beamformed CSI-RS transmission strategies for FD-MIMO. Beamformed CSI-RS uses multiple directional beams so the UE can select and report a preferred beam.

  • Transmission strategies: The standardization process considered extensions of conventional non-precoded CSI-RS and beamformed CSI-RS transmission.These were presented as two CSI-RS transmission strategies.
  • Beamformed CSI-RS: Beamformed CSI-RS adapts the data-transmission weight to the UE-selected beam index in the connected-array architecture.The selected beam’s corresponding weight is used for data transmission.
  • Conventional CSI-RS: Conventional CSI-RS maps TXRUs one-to-one to CSI-RS resources, applies an identical weight to TXRU groups, and lets the UE select a preferred codebook index.The UE chooses the index maximizing channel gain for each subband.
  • Beamformed CSI-RS: Beamformed CSI-RS transmits multiple beams, and the UE feeds back the index of the beam maximizing received power.For rank-1 beamforming, the CSI-RS-port precoder is 1_NB and the data-channel precoder is 1.
  • Feedback implications: Controlling CSI-RS weights can reduce the effective channel dimension and substantially reduce feedback overhead.A beam directed toward a path makes nonmatching spatial signatures approximately orthogonal, yielding one or a few dominant channel taps.

E. CSI feedback mechanisms for FD-MIMO systems

FD-MIMO feedback schemes trade CSI detail against overhead using composite, beam-index, partial, adaptive, and flexible-codebook approaches. Release 13 emphasized composite codebook and beam-index feedback as class-A and class-B ingredients.

  • Standardization focus: Composite codebook and beam-index feedback received attention as main ingredients for class-A and class-B CSI feedback, respectively.Other schemes were deferred to a future release.
  • Composite codebook: Composite codebooks separate vertical and horizontal channel information, with the vertical codebook reported less frequently when vertical angular spread is smaller.A Kronecker product can combine the two codebooks.
  • Beam-index feedback: Beam-index feedback requires multiple beamformed CSI-RSs; rank-one channels need a beam index and CQI, while rank-two dual-polarized channels additionally need co-phase information.The co-phase supports channel orthogonalization between layers.
  • Partial CSI-RS: Partial CSI-RS transmission reduces CSI-RS count from N_H × N_V to N_H + N_V by partitioning the two-dimensional array into horizontal and vertical ports.The eNB reconstructs overall channel information using spatial and temporal correlation among antenna elements.
  • Adaptive CSI feedback: Adaptive CSI feedback combines beamformed and non-precoded CSI-RS by using long-term channel statistics to select beamformed transmission for short-term and subband feedback.The eNB first transmits N_T non-precoded CSI-RSs, then determines spatial direction and transmits beamformed CSI-RSs.
  • Flexible codebook: Flexible codebooks support multiple two-dimensional antenna layouts from one master codebook without increasing the number of codebooks.Examples include (2 × 8), (4 × 4), and (1 × 16) layouts derived from a 16-TXRU master codebook.

IV. PERFORMANCE OF FD-MIMO SYSTEM

System-level evaluations compare FD-MIMO with conventional 8Tx LTE-A under full-buffer and finite-traffic conditions. FD-MIMO provides substantial throughput and spectral-efficiency gains, with the mechanism depending on network loading and antenna spacing.

  • Evaluation setup: FD-MIMO with 16, 32, and 64 transmit antennas was evaluated against conventional 8Tx LTE systems in realistic multicell deployment scenarios.The tested vertical-by-horizontal CSI-RS configurations were 2×8, 4×8, and 8×8.
  • Full-buffer performance: 105% and 484% cell-edge gains were obtained for type I and type II configurations, respectively, over conventional 8Tx LTE-A.The two configurations use different antenna structures and spacing.
  • Antenna spacing: 30% cell-average and 70% cell-edge gains resulted when type II TXRUs were doubled, because sufficient antenna spacing made channel cross-correlation negligible.Type I spectral efficiency did not scale linearly with the number of TXRUs because of insufficient antenna spacing.

V. CONCLUDING REMARKS

The article overviews FD-MIMO standardization in 3GPP LTE-Advanced Pro Release 13, covering key features and major system-design issues. It also identifies unresolved deployment challenges, including pilot overhead reduction, beam adaptation, and advanced channel estimation.

  • V. CONCLUDING REMARKS: The article reviews FD-MIMO standardization in 3GPP LTE-Advanced Pro, emphasizing the discussion and debate during the Release 13 phase.
  • V. CONCLUDING REMARKS: It discusses system-design issues involving channel models, transceiver architectures, pilot transmission, and CSI feedback.
  • V. CONCLUDING REMARKS: FD-MIMO requires features distinct from conventional LTE-A MIMO to use large eNB antenna arrays cost- and space-effectively.
  • V. CONCLUDING REMARKS: The proposed features include array-connected transmitter architecture, beamformed CSI-RS transmissions, and beam-index feedback.
  • V. CONCLUDING REMARKS: Successful future deployment still faces challenges in pilot overhead reduction, beam adaptation and optimization, and channel estimation using time and angular-domain sparsity.
Loading 1601.00019v4…