PaperlayerView exact revision on arXiv

Source-linked AI summary

A Tutorial on Beam Management for 3GPP NR at mmWave Frequencies

Marco Giordani, Michele Polese, Arnab Roy, Douglas Castor, Michele Zorzi

arXiv:1804.01908v2cs.NI

TL;DR

mmWave links require directional beams to overcome propagation losses, but beam alignment complicates initial access and mobility control. The tutorial compares NR beam-management measurements and procedures across architectures and directions, finding trade-offs among accuracy, reactiveness, and overhead that depend on deployment conditions. Its guidelines identify configuration choices for different environments rather than a single universally optimal strategy.

  • Problem

    Severe mmWave propagation conditions require directional links, creating beam-alignment challenges for initial access, tracking, handover, and recovery.

  • Method

    The tutorial evaluates NR beam-management frameworks for initial access and tracking across measurement directions, beamforming configurations, and network settings.

  • Results

    0.6 s to 5.2 s: analog-beamforming initial-access delay when sweeping all directions with SS burst periodicity TSS = 20 ms.

  • Takeaways & Limitations

    Optimal initial-access and tracking designs depend on the deployment environment and involve trade-offs among detection accuracy, reactiveness, and overhead.

Abstract

from arXiv · show

The millimeter wave (mmWave) frequencies offer the availability of huge bandwidths to provide unprecedented data rates to next-generation cellular mobile terminals. However, mmWave links are highly susceptible to rapid channel variations and suffer from severe free-space pathloss and atmospheric absorption. To address these challenges, the base stations and the mobile terminals will use highly directional antennas to achieve sufficient link budget in wide area networks. The consequence is the need for precise alignment of the transmitter and the receiver beams, an operation which may increase the latency of establishing a link, and has important implications for control layer procedures, such as initial access, handover and beam tracking. This tutorial provides an overview of recently proposed measurement techniques for beam and mobility management in mmWave cellular networks, and gives insights into the design of accurate, reactive and robust control schemes suitable for a 3GPP NR cellular network. We will illustrate that the best strategy depends on the specific environment in which the nodes are deployed, and give guidelines to inform the optimal choice as a function of the system parameters.

I. INTRODUCTION

mmWave offers substantially more bandwidth but introduces severe propagation and alignment challenges, making beam management essential for 5G control procedures. This tutorial surveys NR-compatible measurement frameworks and evaluates their accuracy, reactiveness, overhead, and deployment trade-offs.

  • Motivation: mmWave provides much more bandwidth than sub-6 GHz cellular systems but experiences harsher propagation conditions.Higher carrier frequencies create severe pathloss and atmospheric-absorption challenges despite their capacity potential.
  • Motivation: Highly directional phased-array links provide beamforming gain but require precise transmitter–receiver alignment for communication quality.Beam management supports initial access, beam tracking, handover, path selection, and radio-link-failure recovery.
  • Motivation: Beam alignment can delay access procedures and increase performance sensitivity, motivating mmWave-aware extensions to LTE control procedures.Unlike LTE, directional transmissions may be essential before a physical link is established in mmWave systems.
  • Contributions: The tutorial reviews NR beam-management measurement frameworks and evaluates initial-access and tracking strategies under realistic NR settings and channel configurations.The analysis targets fast, accurate, and robust control-plane schemes across different scenarios.
  • Contributions: The study compares downlink/uplink and standalone/non-standalone frameworks, identifying trade-offs involving UE energy, performance, access speed, overhead, robustness, and stability.DL reduces UE energy consumption but may perform worse than UL; SA can favor fast access in stable dense settings, while NSA can reduce overhead and improve robustness.
  • Contributions: Accurate beam management is associated with narrow beams, small subcarrier spacings, denser deployments, frequency diversity, and greater time/frequency resources.Fast access and tracking generally require more allocated resources, increasing overhead; digital or hybrid beamforming can sweep multiple directions simultaneously.

III. FRAME STRUCTURE AND SIGNALS FOR 3GPP NR AT MMWAVE FREQUENCIES

3GPP NR defines a configurable frame structure and synchronization/reference signals that support directional beam management at mmWave frequencies. SS blocks, CSI-RS, and SRS enable downlink and uplink measurements for initial access, mobility, and beam alignment.

  • Frame structure: NR supports mmWave operation with configurable numerologies, OFDM waveforms, and subcarrier spacings of 15 × 2^n kHz for n ≤ 4.Release 15 supports up to 3300 subcarriers, 400 MHz bandwidth, and 10 ms frames.
  • Measurement signals: Beam-management signals are periodically exchanged between transmitters and receivers through downlink and uplink measurement frameworks.Downlink uses gNB-transmitted SS blocks and CSI-RS; uplink uses UE-transmitted SRS.
  • SS blocks: An SS block occupies 4 OFDM symbols and 240 subcarriers and carries the PSS, SSS, PBCH, and PBCH-associated DMRS.The DMRS can support SS-block RSRP estimation.
  • SS blocks: SS blocks form bursts in the first 5 ms, with candidate periodicities TSS ∈ {5, 10, 20, 40, 80, 160} ms and an initial-access assumption of TSS = 20 ms.Above 6 GHz, a burst may contain up to 64 SS blocks, which can map to angular directions.
  • CSI-RS: CSI-RS supports connected-mode RRM measurements by allowing synchronized UEs to search configured resources associated with SS bursts.CSI-RS windows specify periodicity and time/frequency offsets relative to the associated SS burst.
  • Measurement signals: SS- and CSI-based measurements can be combined to identify suitable beam directions across different beamforming architectures.SRS instead monitors uplink channel quality and is scheduled by the serving gNB with specified resources and directions.

IV. BEAM MANAGEMENT FRAMEWORKS FOR 5G CELLULAR SYSTEMS

The tutorial compares beam-management frameworks across standalone/non-standalone architectures and downlink/uplink measurement directions. These frameworks combine sweeping, measurement, beam selection, and reporting to maintain suitable beam pairs over time.

  • Frameworks: The study evaluates three measurement frameworks for initial access and connected-mode tracking.Their main features are summarized comparatively before performance analysis.
  • Beam-management operations: Beam management comprises beam sweeping, beam measurement, beam determination, and beam reporting, repeated periodically to update the optimal beam pair.Measurement evaluates received-signal quality, while determination selects suitable transmitter or receiver beams.
  • Network architectures: Standalone NR operates without LTE control-plane support, whereas non-standalone NR uses an LTE overlay and multi-connectivity.NSA can combine high-rate mmWave links with more robust lower-rate legacy channels and switch paths when links drop.
  • Measurement directions: Downlink frameworks use gNB-transmitted synchronization and reference signals, while uplink frameworks use SRS measurements transmitted by UEs.The uplink is emphasized as increasingly important in heterogeneous cellular networks.

A. Standalone-Downlink (SA-DL) Scheme

The SA-DL scheme uses exhaustive directional scanning and NR reference signals without LTE-overlay support, requiring the UE to determine and report the best beam before random access.

  • Comparison with NSA: Unlike NSA schemes, SA-DL provides no LTE-overlay support, while NSA procedures can use LTE signaling to report directions and schedule access more promptly.The NSA-DL overlay can also report link failure and support fallback to sub-6-GHz connectivity during mmWave recovery.
  • SA-DL procedure: SA-DL performs exhaustive beam sweeping using predefined codebooks that sequentially cover the complete angular space.Both UEs and base stations transmit or receive synchronization and reference signals through codebook directions.
  • SA-DL procedure: Initial access uses SS blocks, while tracking combines SS-burst measurements with CSI-RSs over directions selected according to user needs.CSI-RS directions may cover either the full available direction set or only a subset.
  • SA-DL procedure: The UE selects the beam with maximum SNR when it exceeds a predefined threshold, then uses that sector for subsequent transmissions and receptions.The selected direction provides the resulting antenna gain.
  • SA-DL procedure: After beam determination, the UE waits for the gNB to schedule a direction-specific RACH opportunity for random access and beam reporting.The RACH procedure implicitly informs the serving infrastructure of the UE’s selected direction or directions.
  • Comparison with NSA: NSA-UL measures uplink SRSs through coordinated directional sweeps, sends gNB reports to an LTE coordinator, and uses LTE to report the selected gNB and UE direction.The procedure assumes approximate synchronization and small round-trip propagation times; a 150 m cell corresponds to about 1 µs round-trip delay.
  • Comparison with NSA: The NSA-UL adaptation is not considered by 3GPP and reuses downlink SS-block resources for uplink SRSs.The paper freely adapts the NR frame structure proposed for downlink measurements.

V. PERFORMANCE METRICS AND 3GPP FRAMEWORKS PARAMETERS

The evaluation characterizes beam-management frameworks through detection accuracy, reactiveness, and overhead, while varying NR frame, frequency-diversity, array, deployment, and beamforming parameters.

  • Performance metrics: The three evaluation metrics are misdetection probability, reactiveness, and overhead.These metrics cover detection accuracy, response to access or channel changes, and allocated time-frequency resources.
  • Performance metrics: For initial access, reactiveness is the average time to find the best beam pair; for tracking, it is the time to receive the first CSI-RS after an SS burst.Tracking also considers response time to radio link failure.
  • Frame structure: The NR frame cases use subcarrier spacings of 120 kHz and 240 kHz, with slot durations of 125 µs and 62.5 µs, respectively.The corresponding SS-block bandwidths are 28.8 MHz and 57.6 MHz.
  • Frequency diversity: Frequency diversity either uses remaining SS-symbol bandwidth for data or repeats SS information, with Nrep = 1, 5, or 11 depending on configuration.Repetition uses Nrep = 11 at 120 kHz and Nrep = 5 at 240 kHz.
  • Array geometry: Array geometry determines beamwidth and the number of azimuth and elevation directions, while larger arrays produce narrower beams and higher beamforming gain.The evaluation covers a 120° gNB sector, a 60° elevation range, and UE azimuth scanning over 360°.
  • Beamforming architecture: Analog, hybrid, and digital beamforming trade simultaneous directional coverage, power allocation, flexibility, and hardware complexity.Analog uses one RF chain, hybrid uses KBF RF chains with KBF ≤ M, and digital uses a separate chain and converter per antenna element.
  • Network deployment: The parameter study also varies users per sector and base-station density, using Nuser ∈ {5, 10, 20} and density λb in gNB/km2.These deployment variables are included among the parameters affecting framework performance.

C. Channel Model

The study evaluates detection accuracy through realistic system configurations and Monte Carlo simulations driven by 28 GHz measurements from dense urban New York City deployments.

  • Simulation approach: Detection-accuracy simulations use realistic system-design configurations.The simulations evaluate the probability of misdetection under the study’s configured beam-management parameters.
  • Simulation approach: Results are obtained with a Monte Carlo approach that repeats independent simulations to estimate statistical quantities.The repeated simulations support evaluation of the framework’s detection behavior.
  • Channel model: The channel model is based on recent real-world measurements at 28 GHz in New York City.The setting represents dense urban mmWave microcellular and picocellular networks.
  • Channel model: The deployment includes environmental obstructions such as urban buildings that can occlude transmitter-receiver paths.This models propagation conditions in a dense urban scenario.

VI. RESULTS AND DISCUSSION

The results section evaluates initial access, reactiveness, and overhead for the presented beam-management schemes using the metrics defined earlier.

  • Results scope: The simulations evaluate initial-access detection accuracy through probability of misdetection.They also analyze measurement-framework reactiveness and overhead in subsequent result subsections.
  • Results scope: Reactiveness analysis concerns both network access speed and response to updated channel conditions.The evaluation covers the framework’s ability to respond during mobility and channel variation.
  • Results scope: The study separately analyzes the time-frequency resources allocated to beam-management measurement operations as overhead.The results are organized alongside the detection-accuracy and reactiveness analyses.

A. Detection Accuracy Results

Detection accuracy improves with denser deployments, larger antenna arrays, narrower beams, and frequency diversity, while larger subcarrier spacing can degrade detection unless repetitions mitigate noise.

  • Array size and gNB density: Denser gNB deployments and larger antenna arrays improve detection accuracy by strengthening received power and enabling narrower, higher-gain beams.
  • Array size and gNB density: Misdetection probability decreases monotonically with increasing gNB density or antenna count, whereas omnidirectional UE reception needs 5 gNB/km2 more density than 16-antenna reception for the same PMD.
  • Subcarrier spacing and frequency diversity: With no repetitions, ∆f = 240 kHz gives lower detection accuracy because increased thermal-noise impact degrades the SNR.
  • Subcarrier spacing and frequency diversity: Repeating SS block information across remaining subcarriers improves robustness and mitigates noise during detection.
  • Subcarrier spacing and frequency diversity: The ∆f = 120 kHz configuration without frequency diversity and the ∆f = 240 kHz configuration with Nrep = 5 achieve the same detection accuracy.

B. Reactiveness Results for IA

Initial-access reactiveness is governed by directional-sweep complexity, burst scheduling, beamforming capability, and reporting resources. Faster access generally requires more time/frequency resources or advanced beamforming, increasing overhead or power demands.

  • Sweep complexity: Initial-access delay depends on scanning all gNB and UE directions, while simultaneous beam transmission or reception reduces the number of required sweep steps.
  • Power considerations: Hybrid and digital architectures consume more power than analog architectures at the same ADC resolution, although 3-bit ADC configurations can approach analog power consumption.
  • Number of SS blocks per burst and beamforming technology: Analog beamforming can require 0.6 s with NSS = 64 or 5.2 s with NSS = 8 to scan all directions, making the scheme impractical.
  • Number of SS blocks per burst and beamforming technology: Hybrid or digital UE beamforming increases the configurations able to complete a sweep within one SS block by receiving from multiple directions simultaneously.
  • Beamforming technology: Digital beamforming at the gNB reduces TIA more than at the UE in uplink schemes because the gNB sweeps more narrow-beam directions.
  • SS burst periodicity: Higher SS burst periodicities are unsuitable for mmWave deployment; increasing SS blocks per burst can enable completing the sweep in one burst.
  • Subcarrier spacing: Larger subcarrier spacing shortens OFDM symbols, allowing SS-block transmission in relevant directions to finish earlier.
  • Impact of beam reporting: Beam reporting can have little impact when needed RACH resources fit within one SS burst, leaving sweeping as the dominant initial-access delay.

C. Reactiveness Results for Beam Tracking

Beam-tracking reactiveness depends on CSI-RS scheduling capacity, user and beam counts, periodicities, and the selected scheduling option. The analysis models spatial coverage and the timing available between SS bursts.

  • Tracking metric: Tracking reactiveness is defined as the average time needed to receive the first CSI-RS after each SS burst.
  • User and beam model: The model treats Nuser UEs, each monitoring NCSI,RX directions, as n = NuserNCSI,RX uniformly distributed UEs because LOS conditions can associate them with non-nearest beams.
  • CSI-RS allocation: The average number of CSI-RS transmissions is ZCSI = min{n, k}, where k is the number of available gNB beams.
  • CSI-RS scheduling: Between SS bursts, available CSI-RS time is Ttot,CSI = TSS − Dmax,SS, so scheduling capacity depends on TSS, TCSI, ZCSI, and the selected option.
  • Scheduling options: Option 1 allocates NCSI = ⌊Ttot,CSI/TCSI⌋ CSI-RSs, and one periodicity is insufficient when ZCSI exceeds NCSI.
  • Scheduling options: Option 2 uses an offset OCSI to allocate NCSI = ⌈Ttot,CSI/TCSI⌉ transmissions between consecutive SS bursts.
  • Scheduling caveat: If ZCSI > NCSI, a direction may be received as either an SS block in the next burst or a CSI-RS, depending on scheduling.

TSS + iTCSI + OCSI

Beam-management performance depends strongly on CSI-RS and SS-burst periodicities, architecture, and beamforming choices. Short CSI-RS periodicities improve tracking and neighbor coverage is constrained by CSI-RS scheduling resources, while NSA generally enables faster RLF recovery than SA.

  • CSI-RS periodicity: Small CSI-RS periodicities allow directions to be swept quickly and reduce tracking dependence on the SS-burst periodicity.The tutorial identifies TCSI = 0.625 ms as a small periodicity configuration.
  • Neighbor limits: CSI-RS measurements from neighboring cells cannot overlap in time and frequency, limiting the number of neighbors a gNB can support.The limit depends on available CSI-RS symbols, signal duration, bandwidth scaling ρ, and CSI-RS periodicity.
  • RLF recovery: In SA operation, failure on all monitored directions forces the UE to use SS blocks for recovery or new initial access, preventing data and control transmission meanwhile.Recovery is typically triggered at the subsequent SS burst and is at least subject to the initial-access duration.
  • Uplink versus downlink: Digital UL measurement frameworks can outperform DL tracking in some configurations because gNBs can scan more directions simultaneously, despite higher receiver power consumption.The larger antenna capacity at the gNB enables more simultaneously scanned directions.
  • RLF recovery: SA RLF recovery latency is high across investigated settings and can become unacceptably high in some configurations.The latency is dominated by the initial-access delay.
  • RLF recovery: NSA provides faster RLF recovery than SA across the investigated network configurations by forwarding failure information through the LTE overlay.The LTE link can maintain traffic or trigger recovery while mmWave communication is restored.

D. Overhead Results

Overhead is governed mainly by SS-burst periodicity, subcarrier spacing, repetition strategy, and CSI-RS configuration. CSI-RS overhead is generally small, whereas SS-burst resources dominate total overhead under the evaluated settings.

  • Subcarrier spacing and repetition: With repetitions, Δf = 120 kHz produces higher SS-burst overhead than Δf = 240 kHz because its OFDM symbols last twice as long.The considered repetition counts are 11 for 120 kHz and 5 for 240 kHz.
  • SS-burst overhead: SS-burst overhead follows an inverse relationship with TSS and can reach 43% when TSS = 5 ms with frequency repetitions.For TSS = 20 ms or higher, the overhead is below 10%.
  • CSI-RS overhead: CSI-RS overhead is below 0.008 for TCSI = 5 ms and below 0.06 for TCSI = 0.625 ms.With practical CSI-RS bandwidth below half the total bandwidth, the latter remains below 0.028.
  • Total overhead: The largest contribution to total overhead is the SS-burst term RSS rather than CSI-RS overhead.Subcarrier spacing and repetition strategy produce the main differences in total overhead.
  • Beam reporting: SA beam reporting can require additional overhead because multiple RACH resources may be allocated, whereas NSA uses a single RACH opportunity with overhead 0.0894 · 10^-3.The comparison is reported for beam reporting operations under an SA architecture.

VII. FINAL CONSIDERATIONS

Beam-management performance depends on balancing detection accuracy, reactiveness, and measurement overhead across NR configuration and deployment choices. The preferred design combines parameter settings and architectures according to network density, load, and operating mode.

  • Parameter trade-offs: Smaller subcarrier spacing improves accuracy but increases overhead and can reduce reactiveness.At 120 kHz, the accuracy gain is about 23% for the 4 × 4 array at λ = 30 gNB/km2, while overhead increases by roughly 2 times.
  • Parameter trade-offs: Frequency diversity can improve accuracy substantially, but its overhead and diminishing gains make adoption configuration-dependent.The gain reaches up to 45% for the 4 × 4 array at λ = 60 gNB/km2, while overhead is 5 to 11 times higher; the gain falls to 15% for the 64×4 array.
  • Parameter trade-offs: More SS blocks per burst improve initial-access reactiveness by increasing the chance of completing beam sweeping in one burst, but overhead rises linearly.Burst periodicity also affects reactiveness and overhead, while its impact on initial access disappears when sweeping completes within a single burst.
  • Parameter trade-offs: Shorter CSI-RS periodicity improves tracking reactiveness, especially with many users per gNB, while its overhead remains small relative to SS bursts.Monitoring additional CSI-RSs can avoid network-side increases in transition time or overhead, but may increase UE energy consumption.
  • Deployment and architecture: Higher gNB density improves accuracy and received power but increases interference, neighboring cells, and CSI-RS configuration constraints.Density also enables more users to be served, at the cost of equipment and energy.
  • Deployment and architecture: Digital receiver beamforming improves reactiveness and reduces overhead without penalizing accuracy, whereas omnidirectional reception loses around 30% accuracy at λ = 30 gNB/km2.Hybrid beamforming offers a possible compromise between performance and transceiver complexity or energy consumption.
  • Deployment and architecture: Larger antenna arrays improve accuracy through narrower, higher-gain beams but increase scanning burden and can reduce reactiveness and increase overhead.The MgNB = 4 × MUE = 4 configuration has lower overhead and better reactiveness than MgNB = 64 × MUE = 16 in Fig. 19.
  • Guidelines: A single-burst sweeping and reporting design allows longer TSS, more CSI-RSs for tracking, and lower SS-block overhead.Frequency diversity should depend on gNB load; dense scenarios can use wide SS-burst beams and narrow CSI-RS beams, while standalone suits fast initial access and NSA supports connected-mode resilience.

VIII. CONCLUSIONS

The paper concludes that mmWave beam management must support directional links and updated control procedures for initial access and beam tracking. Its evaluation identifies accuracy, reactiveness, and overhead trade-offs and provides deployment-specific guidelines for selecting strategies.

  • Conclusions: MmWave propagation requires directional beamforming to increase link budget and precise beam-pair tracking for communication.Directional transmission also requires adapting control procedures such as initial access to the absence of an omnidirectional broadcast channel.
  • Conclusions: The tutorial reviews NR beam-management procedures and evaluates how 3GPP-specified parameters affect their performance.It covers standalone and non-standalone architectures, downlink and uplink signaling, and NR settings above 6 GHz.
  • Conclusions: The paper finds trade-offs among detection accuracy, reactiveness, and measurement overhead, so optimal initial-access and tracking strategies depend on deployment conditions and operator needs.The guidelines are intended for different network deployments and specific environments in which nodes are placed.
Loading 1804.01908v2…