Pith. sign in

REVIEW 2 major objections 5 minor 24 references

USRP-Based Single Anchor Positioning: AoA with 5G Uplink Signals, and UWB Ranging

T0 review · 2 major / 5 minor · reviewed 2026-08-12 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read A testbed that combines 5G uplink angle-of-arrival from a single USRP base station with UWB ranging places a walking pedestrian within 10 meters in 88-90% of measurements, and within 4.6 meters at short range.

desk verdict A useful field-testbed paper with a plausible integration of 5G uplink AoA and UWB ranging, but the calibration description is ambiguous enough that the headline accuracy needs a clarifying revision before the numbers can be trusted. read the letter →

arxiv 2411.16496 v1 pith:IJO543ML submitted 2024-11-25 eess.SP cs.SYeess.SY

Thodoris Spanos , Fran Fabra , José A. López-Salcedo , Gonzalo Seco-Granados , Nikos Kanistras , Ivan Lapin , Vassilis Paliouras This is my paper · ORCID
classification eess.SPcs.SYeess.SY
keywords 5GpositioningangleofarrivalUSRPtestbedsoundingreferencesignalUWBrangingsingle-anchorlocalizationpedestrian
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper argues that a single-anchor positioning system built from commercial off-the-shelf software-defined radios can localize a moving pedestrian outdoors without GNSS. The testbed measures the angle of arrival of 5G New Radio uplink sounding reference signals with a three-element antenna array on an Ettus N310, and fuses that angle with ultra-wideband ranging to the base station. The authors report that 88% of position estimates at 2.4 GHz and 90% at 5.8 GHz are within 10 meters of ground truth over a trajectory reaching 90 meters, and that for distances under 40 meters the 90th-percentile error drops to 4.6 meters and 3.4 meters respectively. If these numbers hold, single-node positioning becomes a practical complement to GNSS in indoor and urban canyon settings.

What carries the argument

The load-bearing mechanism is the phase-alignment calibration for the Ettus N310's two independent local-oscillator domains. Because the two daughterboards each serve a pair of receive channels with a random phase offset between runs, the testbed injects a common continuous tone through a 1-2 splitter into one channel of each pair, estimates the differential phase by cross-correlation, and compensates it in real time. This makes three of the four receive channels coherent so that a three-element uniform linear array can feed ESPRIT for angle-of-arrival estimation; a snapshot capture scheme keeps the large IQ files manageable, and UWB ranging provides the distance that turns the angle into a position.

What would settle it

Measure the calibration phase offset immediately before and after a pedestrian run; if the offset shifts by more than a small fraction of a wavelength, or if rerunning the same trajectory with the test tone switched off changes the reported angles, the calibration assumption fails.

Watch

Extended reading notes

Core claim

The central discovery is that a single base station can track a pedestrian by combining two complementary measurements: an angular estimate from 5G uplink SRS, obtained with a three-element uniform linear array and the ESPRIT algorithm, and a distance estimate from UWB two-way ranging. The enabling step is a real-time calibration of the Ettus N310 receiver, which injects a common tone into one channel of each daughterboard pair to measure and remove the phase offset caused by the two separate local oscillators. With this alignment, the testbed achieves sub-10-meter errors in 88% of cases at 2.4 GHz and 90% at 5.8 GHz, and 90th-percentile errors below 4.6 meters and 3.4 meters respectively for ranges under 40 meters.

Load-bearing premise

The positioning accuracy depends on the assumption that the phase difference between the two receiver channel pairs, measured at setup by injecting a common test tone, stays unchanged throughout the walk and does not contaminate the 5G signal measurement.

Editorial extensions

If this is right

  • A single base station, using only uplink 5G signals and UWB ranging, can provide outdoor pedestrian localization where GNSS is unavailable or degraded.
  • The sub-4.6-meter 90th-percentile error below 40 meters is accurate enough for pedestrian navigation and asset tracking applications in that range.
  • The phase-calibration trick removes the need for an external clock or a four-channel fully synchronized array, lowering the hardware cost of coherent angle-of-arrival estimation.
  • The 5.8 GHz band gives slightly better accuracy than 2.4 GHz, suggesting that higher-frequency 5G bands are favorable for single-anchor positioning.
  • The snapshot capture method keeps storage and processing feasible at up to 61.44 MSps without sacrificing positioning accuracy.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Beyond the paper: the same calibration and ESPRIT pipeline could be applied to the vehicular configurations in Table I; the open question is whether the phase alignment remains stable under Doppler and vibration.
  • Beyond the paper: a two-dimensional array or adding elevation estimation could lift the system to 3D positioning, especially because the barometer readings were too unstable to use.
  • Beyond the paper: reporting the calibration phase offset measured before and after each run would provide a direct check on the stability assumption, and would make the accuracy numbers reproducible.
  • Beyond the paper: the decreasing UWB reliability beyond 40 meters suggests that a fusion with 5G timing-based ranging could extend the reliable range without new hardware.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

2 major / 5 minor

Summary. The paper describes a single-anchor positioning testbed that combines an Ettus E312 transmitter, an Ettus N310 receiver with a three-element uniform linear array for angle-of-arrival estimation from 5G NR uplink sounding reference signals, and ultra-wideband ranging. Field trials with a pedestrian user at 2.4 GHz and 5.8 GHz produce position-error cumulative distribution functions; the authors report sub-10 m accuracy in 88% and 90% of cases, respectively, and sub-4.6 m / sub-3.4 m errors at the 90th percentile for distances below 40 m. The manuscript also describes a real-time phase-calibration procedure for the N310's two daughterboard channel pairs, snapshot-based IQ capture, and an MQTT-based control architecture.

Significance. If the reported measurements are reliable, the paper contributes a useful and relatively rare experimental testbed for 5G single-anchor positioning, combining uplink SRS-based AoA with UWB ranging in outdoor pedestrian conditions. Strengths include the use of real hardware, independent laser-based ground-truth landmarks, and the absence of any fitting loop between the measured angles/ranges and the final position errors. However, the central quantitative claims are currently undermined by an ambiguity in the RF calibration routing and by missing-data issues in the CDF computation; these need to be resolved before the accuracy percentages can be taken at face value.

major comments (2)
  1. [Section III.A (with Section II.C.5)] The real-time calibration procedure as written is incompatible with the stated three-element ULA. Section II.C.5 says that three of the four Rx channels feed a three-element ULA for AoA estimation. Section III.A then states that a common signal is injected into one channel of each daughterboard pair via a 1-2 splitter. That injection occupies two channels, leaving only two channels available for SRS reception, not three. If instead one of the calibration-injected channels is also used as an array element, then the calibration tone is superimposed on the SRS, and the paper does not describe any frequency offset, filtering, or time-duplexing that would separate the two. Either reading changes the array size or element quality used for AoA, and both could materially affect the position-error statistics in Fig. 6 and the headline 88%/90% accuracy claims. Please provide a per-channel RF routing diagram and state explicitly which physical Rx channels carry the ULA elements, which carry the calibration tone, and how the calibration signal and SRS are separated in frequency or time.
  2. [Section IV-C and Section IV-B] The reported CDFs are computed on a subset of the trajectory. The text states that beyond 40 m the UWB ranging becomes less reliable and 'occasionally fail[s] to provide a position estimate every second', with the failure rate increasing with distance. The percentages 'sub-10 meter accuracy in 88% and 90% of cases' are therefore conditional on successful UWB ranging, not representative of the full trajectory. The paper should report the total number of position estimates, the number and spatio-temporal distribution of dropped UWB measurements, and provide either complete-case and all-track CDFs or distance-stratified results. In addition, ground truth is established by interpolation between stationary landmarks (Section IV-B), which can smooth out genuine errors between landmarks; an estimate of the interpolation error or a continuously surveyed reference would strengthen the validation. No trial counts or error bars are provided for the reported percentages, so the uncertainty of the 88%/90% figures is unknown.
minor comments (5)
  1. [Fig. 6] The y-axis label 'Cumulative Density Function' should be 'Cumulative Distribution Function'.
  2. [Section III.D, step 7] The use of the LCMV beamformer to select among multiple AoA candidates when the estimated channel order is greater than one is not explained; the selection rule should be stated.
  3. [Section IV.A] The trajectory includes points up to 90 m, but all quantitative accuracy claims are separated only by the 40 m threshold; a distance-binned error table would help the reader understand where the errors occur.
  4. [Section IV.C] The paper does not state how many independent pedestrian runs were performed; if the CDFs are from a single run, this should be stated explicitly.
  5. [Section III.A] The abstract emphasizes 'real-time calibration', but it is not clear whether the calibration is repeated during the pedestrian motion or only once at startup; please clarify the calibration schedule.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the reported localization accuracy is an experimental result checked against independent laser-based ground truth, and the self-citations are not load-bearing.

full rationale

The central claim is empirical rather than derived: AoA is estimated from SRS snapshots received by the N310 ULA and range is obtained from UWB two-way ranging, and the resulting position estimates are compared with a ground-truth trajectory established by laser measurements at surveyed landmarks. None of the paper's equations or processing steps defines the reported error metric in terms of the inputs used to produce it, and no fitted parameter is renamed as a prediction; in particular, the 88%/90% sub-10 m and 4.6 m/3.4 m CDF statements are summary statistics of measured position errors, not outputs of a model fitted to those same errors. The paper does cite two prior works by overlapping authors, [14] and [15], to support the N310 two-daughterboard phase-offset limitation and the prior AoA algorithm study, but these citations support an ancillary calibration motivation rather than the numerical accuracy claim, and the calibration procedure itself is described in the paper and applied before comparison with independent ground truth. The possible ambiguity in how the 1-2 splitter calibration tone coexists with the three-element ULA is a hardware-description concern, not a circularity, because it does not make the measured position errors equal their own inputs. No circular step is therefore present.

Assumptions & free parameters 0 free parameters · 5 assumptions · 0 invented entities

No explicit free parameters are fitted to achieve the reported accuracy; phase offsets are measured from a calibration tone and distances are measured by UWB. The central claim rests mainly on domain assumptions about the array model, calibration validity, ground truth construction, and the subset of measurements that survive UWB dropout, all listed below.

assumptions (5)
  • domain assumption Far-field plane-wave signal model for the three-element uniform linear array.
    Invoked implicitly by the MUSIC/ESPRIT/AoA processing in Section III-D step 6; near-field or strongly curved wavefronts would bias angle estimates.
  • domain assumption The real-time injected calibration tone accurately measures the phase offset between N310 daughterboard channel pairs and does not corrupt the SRS reception.
    Section III-A describes the 1-2 splitter calibration during normal operation, but the paper does not verify that the offset stays constant over a dynamic run or that the tone is fully isolated from the AoA signal path.
  • domain assumption Ground truth is accurately reconstructed by linear interpolation between stationary landmarks measured with a laser and synchronized by UWB timestamps.
    Section IV-B constructs the reference trajectory this way; any interpolation or synchronization error is attributed to the testbed as position error.
  • domain assumption The UWB range measurements are available and accurate at the distances and times used in the CDF.
    Section IV-C reports that UWB becomes unreliable beyond 40 m and occasionally produces no estimate, so the CDF is computed on a subset of the trajectory.
  • domain assumption A stored replica of the transmitted SRS is known to the receiver for timing synchronization.
    Section III-D step 3 cross-correlates the captured signal with a replica; this assumes the transmitted waveform is known and unchanged.

how reviews work

0 comments
Cite this review

Pith. "Pith review of USRP-Based Single Anchor Positioning: AoA with 5G Uplink Signals, and UWB Ranging." pith.science (2026). https://pith.science/paper/IJO543ML

@misc{pith2026241116496,
  author       = {Pith},
  title        = {Pith review of: USRP-Based Single Anchor Positioning: AoA with 5G Uplink Signals, and UWB Ranging},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/IJO543ML}},
  note         = {Machine review of arXiv:2411.16496}
}
read the original abstract

This paper presents a novel testbed designed for 5th-Generation (5G) positioning using Universal Software Radio Peripherals (USRPs). The testbed integrates multiple units: an Operation Unit for test management, a User Unit equipped with an Ettus E312 USRP, and a Station Unit featuring an Ettus N310 USRP equipped with a three-element Uniform Linear Array for Angle of Arrival estimation. Alongside ultra wideband ranging, the testbed estimates the user's position relative to the base station. Signal processing algorithms are executed in a dedicated processing unit. Key challenges addressed include phase misalignment between RX channel pairs due to different Local Oscillators in the Ettus N310, necessitating real-time calibration for precise signal alignment. High sampling rates (up to 61.44 MSps) result in large IQ sample files, managed efficiently using a snapshot technique to optimize storage without compromising testbed positioning capabilities. The testbed synchronizes angular measurements with ranging estimates allowing consistent performance evaluation for real-life cases of dynamic users (e.g. pedestrian). Experimental results demonstrate the testbed's effectiveness in achieving accurate pedestrian user localization.

Figures

Figures reproduced from arXiv: 2411.16496 by the authors.

Figure 1
Figure 1. Block diagram presenting the testbed architecture. Operation and Processing Units are hosted in the Station Unit PC. [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Communication between the different units of the testbed. [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 4
Figure 4. , starts at point A1 (15 m from the station unit antenna), proceeds through points A2, T3, T2, T1, and returns to A1, with a maximum distance of 90 m at T2. The pedestrian user moves at 4 km/h, remaining stationary for 30 s at each point. B. Ground Truth The landmarks along the trajectory were precisely measured relative to the base station using a laser measure, providing known reference points. The high accuracy o… view at source ↗
Figures from the paper (2 more)
Figure 5
Figure 5. Figure 5: Estimated position for 2.4 GHz (left) and 5.8 GHz (right). [PITH_FULL_IMAGE:figures/full_fig_p004_5.png]
Figure 6
Figure 6. Figure 6: Position Error CDF for 2.4 GHz (left) and 5.8 GHz (right). [PITH_FULL_IMAGE:figures/full_fig_p004_6.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

24 extracted references · 20 canonical work pages

  1. [1]

    Millimeter- Wave Mobile Sensing and Environment Mapping: Models, Algorithms and Validation,

    C. Baquero Barneto, E. Rastorgueva-Foi, M. F. Keskin, T. Riihonen, M. Turunen, J. Talvitie, H. Wymeersch, and M. Valkama, “Millimeter- Wave Mobile Sensing and Environment Mapping: Models, Algorithms and Validation,” IEEE Transactions on Vehicular Technology , vol. 71, no. 4, pp. 3900–3916, 2022

    work page 2022

  2. [2]

    Invited Paper: Challenges and Opportu- nities in Enabling Secure 5G Positioning,

    A. K. Dutta and M. Singh, “Invited Paper: Challenges and Opportu- nities in Enabling Secure 5G Positioning,” in 2023 15th International Conference on COMmunication Systems & NETworkS (COMSNETS) , 2023, pp. 498–504

    work page 2023

  3. [3]

    Soft Information-Based Localization for 5G Networks and Beyond,

    F. Morselli, S. Modarres Razavi, M. Z. Win, and A. Conti, “Soft Information-Based Localization for 5G Networks and Beyond,” IEEE Transactions on Wireless Communications , vol. 22, no. 12, pp. 9923– 9938, 2023

    work page 2023

  4. [4]

    5G Communication Signal Based Localization with a Single Base Station,

    Y . Li, Z. Zhang, L. Wu, J. Dang, and P. Liu, “5G Communication Signal Based Localization with a Single Base Station,” in 2020 IEEE 92nd Vehicular Technology Conference (VTC2020-Fall), 2020, pp. 1–5

    work page 2020

  5. [5]

    A Comparative Study of 3D UE Positioning in 5G New Radio with a Single Station,

    B. Sun, B. Tan, W. Wang, and E. S. Lohan, “A Comparative Study of 3D UE Positioning in 5G New Radio with a Single Station,” Sensors, vol. 21, no. 4, 2021. [Online]. Available: https: //www.mdpi.com/1424-8220/21/4/1178

    work page 2021

  6. [6]

    High- Precision Single Base Station Localization Assisted by Beamform- ing,

    Y . Liu, Z. Gong, Z. Zhang, L. Wu, R. Gui, J. Dang, and B. Zhu, “High- Precision Single Base Station Localization Assisted by Beamform- ing,” in 2021 2nd Information Communication Technologies Conference (ICTC), 2021, pp. 178–183

    work page 2021

  7. [7]

    Multipath Assisted Single Base Station Positioning for NLOS Environment,

    L. Xie, Y . Zhang, Y . Wang, W. Nie, and M. Zhou, “Multipath Assisted Single Base Station Positioning for NLOS Environment,” in 2022 IEEE International Symposium on Antennas and Propagation and USNC- URSI Radio Science Meeting (AP-S/URSI) , 2022, pp. 307–308

    work page 2022

  8. [8]

    Experimental Evaluation of AoA Algorithms using NI USRP Software Defined Radios,

    B. Rares, C. Codau, A. Pastrav, T. Palade, H. Hedesiu, B. Balauta, and E. Puschita, “Experimental Evaluation of AoA Algorithms using NI USRP Software Defined Radios,” in 2018 17th RoEduNet Conference: Networking in Education and Research (RoEduNet) , 2018, pp. 1–6

    work page 2018

Show all 24 references
  1. [9]

    Performance Evaluation of Single Base Station ToA-AoA Localization in an LTE Testbed,

    A. Blanco, N. Ludant, P. J. Mateo, Z. Shi, Y . Wang, and J. Widmer, “Performance Evaluation of Single Base Station ToA-AoA Localization in an LTE Testbed,” in2019 IEEE 30th Annual International Symposium on Personal, Indoor and Mobile Radio Communications (PIMRC), 2019, pp. 1–6

    2019

  2. [10]

    Performance Evaluation of E-CID based Positioning on OAI 5G-NR Testbed,

    D. Li, X. Chu, L. Wang, Z. Lu, S. Zhou, and X. Wen, “Performance Evaluation of E-CID based Positioning on OAI 5G-NR Testbed,” in 2022 IEEE/CIC International Conference on Communications in China (ICCC), 2022, pp. 832–837

    2022

  3. [11]

    Experimental Validation of Single Base Station 5G mm Wave Positioning: Initial Findings,

    Y . Ge, H. Chen, F. Jiang, M. Zhu, H. Khosravi, S. Lindberg, H. Her- bertsson, O. Eriksson, O. Brunneg ˚ard, B.-E. Olsson, P. Hammarberg, F. Tufvesson, L. Svensson, and H. Wymeersch, “Experimental Validation of Single Base Station 5G mm Wave Positioning: Initial Findings,” in ...

    2022

  4. [12]

    3GPP TS 38.211: NR Physical channels and modulation (Re- lease 18),

    3GPP, “3GPP TS 38.211: NR Physical channels and modulation (Re- lease 18),” 3rd Generation Partnership Project, Technical Specification Group Radio Access Network , 2024

    2024

  5. [13]

    [Online]

    (2024) Grove - Barometer (High-Accuracy). [Online]. Available: https://wiki.seeedstudio.com/Grove-Barometer-High-Accuracy/

    2024

  6. [14]

    Pathway to Coherent Phase Acquisition in Multi-Channel USRP SDRs for Direction of Arrival Estimation,

    A. Xhafa, F. Fabra, J. A. L ´opez-Salcedo, G. Seco-Granados, and I. Lapin, “Pathway to Coherent Phase Acquisition in Multi-Channel USRP SDRs for Direction of Arrival Estimation,” in 2023 Tenth In- ternational Conference on Software Defined Systems (SDS) , 2023, pp. 1–8

    2023

  7. [15]

    Angle of Arrival Estimation Using SRS in 5G NR Uplink Scenarios,

    T. Spanos, F. Fabra, J. A. L ´opez-Salcedo, G. Seco-Granados, N. Kanistras, I. Lapin, and V . Paliouras, “Angle of Arrival Estimation Using SRS in 5G NR Uplink Scenarios,” in WIPHAL, 2024. [Online]. Available: https://api.semanticscholar.org/CorpusID:271431302

    2024

  8. [16]

    Synchronization Algorithms for OFDM Systems (IEEE802.11a, DVB-T): Analysis, Simulations, Optimization and Im- plementation Aspects,

    C. E. Hajjar, “Synchronization Algorithms for OFDM Systems (IEEE802.11a, DVB-T): Analysis, Simulations, Optimization and Im- plementation Aspects,” Ph.D. dissertation, Dec. 2007, phD thesis

    2007

  9. [17]

    Carrier frequency offset estimation in OFDM systems,

    P. K. Nishad and P. Singh, “Carrier frequency offset estimation in OFDM systems,” in 2013 IEEE Conference on Information & Communication Technologies, 2013, pp. 885–889

    2013

  10. [18]

    A new look at the statistical model identification,

    H. Akaike, “A new look at the statistical model identification,” IEEE Transactions on Automatic Control , vol. 19, no. 6, pp. 716–723, 1974

    1974

  11. [19]

    Multiple emitter location and signal parameter estimation,

    R. Schmidt, “Multiple emitter location and signal parameter estimation,” IEEE Transactions on Antennas and Propagation , vol. 34, no. 3, pp. 276–280, 1986

    1986

  12. [20]

    ESPRIT-estimation of signal parameters via ro- tational invariance techniques,

    R. Roy and T. Kailath, “ESPRIT-estimation of signal parameters via ro- tational invariance techniques,”IEEE Transactions on Acoustics, Speech, and Signal Processing , vol. 37, no. 7, pp. 984–995, 1989

    1989

  13. [21]

    Joint angle and delay estimation using shift-invariance techniques,

    A.-J. van der Veen, M. Vanderveen, and A. Paulraj, “Joint angle and delay estimation using shift-invariance techniques,” IEEE Transactions on Signal Processing , vol. 46, no. 2, pp. 405–418, 1998

    1998

  14. [22]

    Improving the resolution performance of eigenstructure- based direction-finding algorithms,

    A. Barabell, “Improving the resolution performance of eigenstructure- based direction-finding algorithms,” in ICASSP ’83. IEEE International Conference on Acoustics, Speech, and Signal Processing , vol. 8, 1983, pp. 336–339

    1983

  15. [23]

    High-resolution frequency-wavenumber spectrum analysis,

    J. Capon, “High-resolution frequency-wavenumber spectrum analysis,” Proceedings of the IEEE , vol. 57, no. 8, pp. 1408–1418, 1969

    1969

  16. [24]

    An algorithm for linearly constrained adaptive array process- ing,

    O. Frost, “An algorithm for linearly constrained adaptive array process- ing,” Proceedings of the IEEE , vol. 60, no. 8, pp. 926–935, 1972. 5

    1972

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.