Ultra-Wideband Double-Directional Channel Measurements and Statistical Modeling in Urban Microcellular Environments for the Upper-Midband/FR3

Basic Information

• Paper ID: 2412.20755
• Title: Ultra-Wideband Double-Directional Channel Measurements and Statistical Modeling in Urban Microcellular Environments for the Upper-Midband/FR3
• Authors: Naveed A. Abbasi, Kelvin Arana, Siddhant Singh, Atulya Bist, Vikram Vasudevan, Tathagat Pal, Jorge Gomez-Ponce, Young-Han Nam, Charlie Zhang, Andreas F. Molisch
• Classification: eess.SY cs.SY
• Publication Date: December 31, 2024 (arXiv v2)
• Paper Link: https://arxiv.org/abs/2412.20755

Abstract

The upper-midband spectrum (FR3, 6-24 GHz) is increasingly important for next-generation wireless networks. This paper presents the first ultra-wideband double-directional channel measurements in urban microcellular environments across most of the FR3 spectrum (6-14 GHz). The study analyzed over 25,000 directional power delay profiles, providing deep insights into line-of-sight (LoS) and obstructed line-of-sight (OLoS) conditions. Subsequently, statistical modeling was performed on path loss, shadowing effects, delay spread, and angular spread. As the first ultra-wideband double-directional measurement campaign in this frequency range, it provides critical insights for spectrum allocation, channel modeling, and advanced communication system design.

Research Background and Motivation

Problem Definition

1. Growing Spectrum Demand: Mobile data traffic is projected to increase 2.5-fold from 2024 to 2030, reaching 300 EB per month, necessitating new spectrum resources
2. FR3 Spectrum Gap: The upper-midband spectrum (6-24 GHz) is defined as FR3 by 3GPP, but lacks detailed propagation characteristic studies
3. Measurement Technology Limitations: Existing measurements are predominantly narrowband or unidirectional, lacking ultra-wideband double-directional measurement data

Research Significance

• System Design Foundation: Path loss and shadowing effects are fundamental requirements for assessing system coverage and reliability
• Spectrum Management: Provides evidence for spectrum regulatory authorities to determine the most suitable spectrum portions for cellular services
• Channel Modeling: Current 3GPP 38.901 models are based solely on measurement data below 6 GHz and above 24 GHz

Limitations of Existing Work

• Most FR3 measurements employ SISO configurations, unable to provide directional information required by multi-antenna systems
• Existing double-directional measurements are primarily conducted in indoor environments or short-distance scenarios
• Lack of ultra-wideband measurements covering most of the FR3 spectrum

Core Contributions

1. First Ultra-Wideband Double-Directional Measurements: Conducted the first ultra-wideband double-directional channel measurements in urban microcellular environments in the 6-14 GHz band
2. Large-Scale Data Analysis: Analyzed over 25,000 directional power delay profiles covering transmission distances of 60-400 meters
3. OLoS Propagation Characteristics: Focused analysis on OLoS propagation conditions caused by vegetation obstruction, demonstrating the need to treat it as an independent category
4. Statistical Modeling: Provided statistical characteristics and frequency-dependent analysis for path loss, delay spread, and angular spread

Methodology Details

Measurement System Architecture

Hardware Configuration:

• Vector Network Analyzer (VNA): Covering 6-14 GHz, 8001 frequency points, 1 MHz spacing
• Horn Antennas: HGHA618 model, one each at transmitter and receiver
• Radio Frequency over Fiber (RFoF): Connecting receiver antenna and VNA
• High-precision mechanical positioners: Enabling double-directional angular scanning

Measurement Parameters:

• Transmitter height: 20.38 m
• Receiver height: 1.7 m
• Transmitter azimuth range: -60°, 60°, 10° step
• Receiver azimuth range: 0°, 360°, 10° step
• Receiver elevation angle range: -20°, 20°, 10° step

Measurement Scenarios

Environment Description:

• Location: University of Southern California Los Angeles Campus
• Scenario: Urban microcellular with mixed buildings, streets, and open spaces
• Measurement Points: 11 receiver locations, distances 65.1-436.1 meters
• Propagation Conditions: 1 LoS point, 10 OLoS points (vegetation-obstructed)

Data Processing Methods

Channel Transfer Function Calibration:

H(f, φTx, φRx, θ̃Rx; d) = Hmeas(f, φTx, φRx, θ̃Rx; d) / HOTA(f)

Power Delay Profile Calculation:

Pcalc(τ, φTx, φRx, θ̃Rx, d) = |F^(-1){H(f, φTx, φRx, θ̃Rx, d)}|²

Noise Suppression: Employing threshold processing and delay gating with 22 dB dynamic range

Omnidirectional PDP Construction:

P'omni(τ; d) = max(φTx,φRx) Σk P(τ, φTx, φRx, θ̃k; d)

Parameter Extraction

Path Loss Modeling:

PLdB(d) = α + 10β log10(d) + ε

where ε ~ N(0,σ) represents shadowing effects

RMS Delay Spread:

στ = √[(∫τ Pi(τ)τ²dτ / ∫τ Pi(τ)dτ) - (∫τ Pi(τ)τdτ / ∫τ Pi(τ)dτ)²]

Angular Spread: Using Fleury definition

σ° = √[Σφ |e^(jφ) - μφ|² APSo(φ) / Σφ APSo(φ)]

Experimental Setup

Measurement Configuration

• Frequency Range: 6-14 GHz, divided into eight 1 GHz subbands
• Measurement Time: Primarily conducted at night to reduce environmental interference
• Calibration Method: Daily over-the-air calibration at 56.45-meter reference point
• Data Volume: 2340 transfer functions generated per transmitter-receiver location pair

Evaluation Metrics

• Path Loss: Deviation from free-space loss
• Shadowing Effects: Standard deviation of Gaussian distribution fitting
• RMS Delay Spread: dB·s on logarithmic scale
• Angular Spread: Dimensionless Fleury definition value 0,1

Analysis Methods

• Linear regression with 95% confidence intervals
• Gaussian fitting of cumulative distribution functions
• Frequency-dependent analysis
• Distance-dependent modeling

Experimental Results

Power Delay Profile Analysis

LoS Condition (Rx1, d=65.1m):

• 6-7 GHz: Received power -84.88 dB, close to Friis model prediction (-84.97 dB)
• 13-14 GHz: Received power -91.54 dB, close to Friis model prediction (-91.32 dB)
• LoS component 25-35 dB stronger than other components

OLoS Condition (Rx5, d=143.6m):

• 6-7 GHz: Received power -116.87 dB, 25 dB below Friis prediction
• 13-14 GHz: Received power -126.76 dB, 28.5 dB below Friis prediction
• Vegetation loss causes LoS component and reflected component power to be comparable

Path Loss Modeling Results

Omnidirectional Case:

• Path loss exponent β: 3.87-4.57 (theoretical value 2)
• Shadowing effect standard deviation: 2.60-4.67 dB
• Higher frequencies exhibit greater path loss

Strongest Beam Case:

• Path loss exponent β: 3.64-4.33
• Shadowing effect standard deviation: 4.74-6.22 dB
• Greater shadowing effect variance compared to omnidirectional case

Delay Spread Characteristics

Statistical Features:

• Omnidirectional RMS delay spread mean: -84.54 dB·s, standard deviation 8.59 dB·s
• Strongest beam RMS delay spread mean: -89.77 dB·s, standard deviation 6.80 dB·s
• Minimal variation across frequency bands, indicating weak frequency dependence

Distance Dependence:

• Weak correlation with distance, β values span zero
• Primarily influenced by location-specific scattering and multipath environment

Angular Spread Analysis

Transmitter Angular Spread:

• Decreases with increasing distance, β values -0.05 to -0.02
• Mean range: 0.19-0.21, weak frequency dependence

Receiver Angular Spread:

• Azimuth spread larger than transmitter (360° coverage vs 120°)
• Elevation spread smaller (measurement range limitation)
• Inconsistent distance dependence, wide confidence intervals

Related Work

FR3 Spectrum Measurement Status

• SISO Measurements: Predominantly narrowband or wideband measurements, lacking directional information
• Unidirectional Measurements: Some work provides single-end directional resolution, but insufficient
• Double-Directional Measurements: Primarily limited to indoor environments or short-distance scenarios

Advantages Compared to This Work

1. Spectrum Coverage: First ultra-wideband measurement covering 6-14 GHz
2. Environmental Realism: Urban microcellular environment, practical deployment scenario
3. Data Scale: 25,000+ directional PDPs, large data volume
4. Analysis Depth: Multi-parameter statistical modeling and frequency-dependent analysis

Conclusions and Discussion

Main Conclusions

1. Significant Vegetation Impact: Additional loss caused by vegetation in OLoS conditions requires separate modeling
2. Frequency Dependence: Path loss increases with frequency, but delay spread and angular spread show weak frequency dependence
3. Distance Dependence: Path loss strongly distance-dependent, delay spread and angular spread weakly distance-dependent
4. Multipath Characteristics: Relative importance of reflected components increases in OLoS conditions

Limitations

1. Limited Measurement Points: Only 11 locations, relatively limited statistical samples
2. Single Scenario: Covers only one type of urban microcellular environment
3. Selection Bias: Measurement point selection ensures reasonable received power, potentially affecting statistical representativeness
4. Time Cost: Mechanical scanning is time-consuming, limiting measurement scale

Future Directions

1. Expanded Measurement Scale: More locations and environment types
2. Electronic Scanning: Adopt electronic beamforming to reduce measurement time
3. Unbiased Measurements: Regular grid measurements to assess coverage probability
4. Other Frequency Bands: Extension to complete FR3 spectrum (6-24 GHz)

In-Depth Evaluation

Strengths

1. Pioneering Work: First FR3 ultra-wideband double-directional measurements, filling a field gap
2. Rigorous Methodology: Well-designed measurement system, scientifically sound data processing methods
3. Comprehensive Analysis: Multi-dimensional parameter analysis, complete statistical modeling
4. Practical Value: Provides important reference data for 5G/6G system design

Shortcomings

1. Sample Limitations: Limited measurement environments and locations, generalizability to be verified
2. Theoretical Analysis: Lacks in-depth theoretical explanation of observed phenomena
3. Comparative Validation: Insufficient detailed comparison with existing channel models
4. Design Guidance: Limited specific guidance for system design

Impact

1. Academic Contribution: Provides first measured data for FR3 channel modeling
2. Standardization Value: Can serve as reference for standardization organizations such as 3GPP
3. Engineering Application: Provides propagation characteristic foundation for FR3 system deployment
4. Subsequent Research: Establishes foundation for follow-up research in this field

Applicable Scenarios

1. System Design: Coverage planning for FR3 cellular systems
2. Spectrum Management: Spectrum allocation decisions by regulatory authorities
3. Channel Modeling: Refinement of standard channel models such as 3GPP
4. Academic Research: Reference benchmark for related propagation characteristic research

References

The paper cites abundant related work, primarily including:

• 3GPP 38.901 channel model standards
• Various measurement work on FR3 spectrum
• Classical literature on UWB propagation measurements
• Channel parameter extraction and modeling methods

Overall Assessment: This is a high-quality measurement research paper with pioneering significance in FR3 spectrum channel characteristic research. Despite limitations in measurement scale and environmental coverage, its rigorous methodology and detailed data provide important foundational contributions to the field.