Fuse-then-Detect for Passive UAV Localization Using Multi-UE 5G Uplink Signals

Charlie Jianzhong Zhang; Murat Bayraktar; Nuria Gonz\'alez-Prelcic; Vishnu Ratnam; Wenyu Huang

REVIEW 3 major objections 2 minor 2 cited by

Fuse-then-Detect for Passive UAV Localization Using Multi-UE 5G Uplink Signals

T0 review · 3 major / 2 minor · reviewed 2026-07-15 · grok-4.5

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

Pith's one-line read Passive UAV localization from multi-UE 5G uplink SRS reaches 4.84 m median 3D error after LOS-referenced synchronization.

desk verdict Abstract-only claim of first multi-UE uplink 5G SRS passive UAV localization with sub-ns sync and 4.84 m median error; idea is clean but completely uncheckable. read the letter →

arxiv 2607.11955 v1 pith:RG4XAFBY submitted 2026-07-12 eess.SP cs.SYeess.SY

Wenyu Huang , Nuria González-Prelcic , Vishnu Ratnam , Murat Bayraktar , Charlie Jianzhong Zhang This is my paper · ORCID

classification eess.SPcs.SYeess.SY

keywords passiveUAVlocalization5GuplinksensingsoundingreferencesignalISACmulti-UEfusionLOS-referencedsynchronizationbistaticgeometryFR1

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 claims that base stations can locate low-altitude UAVs passively by listening to 5G sounding-reference-signal (SRS) pilots that ordinary user phones already transmit, without any dedicated radar or extra downlink sensing. The difficulty is that each phone has its own clock and timing loop, so residual timing, frequency, and amplitude errors scramble the weak UAV echo, and a single phone’s return is buried under line-of-sight and urban clutter. The authors show that those residuals can be stripped to sub-nanosecond accuracy by reusing the existing timing-advance command and a conjugate product of adjacent SRS occasions, after which a joint detector that fuses all phones into a shared three-dimensional search space can accumulate a normalized geometric contrast strong enough for reliable detection. In a cluttered urban FR1 simulation with four pedestrian users and 100 MHz New Radio bandwidth the pipeline yields 4.84 m median 3D position error. If the approach works in the field, existing cellular infrastructure becomes a dense, always-on sensor for low-altitude airspace safety without new spectrum or hardware.

What carries the argument

LOS-referenced synchronization followed by a fuse-then-detect joint detector: the former cancels residual timing/frequency/amplitude impairments with the TA command and conjugate products; the latter searches a common 3D UAV state space and accumulates a geometry-normalized contrast across UEs so that weak multi-path echoes become detectable against LOS and clutter.

What would settle it

A field or ray-traced trial with real multi-UE 5G SRS in which the residual timing after the proposed LOS-referenced correction exceeds a few nanoseconds, or the fused multi-UE detector fails to produce a median 3D error near 5 m under comparable clutter and SNR.

Watch

Extended reading notes

Core claim

The first uplink passive-sensing framework for UAV localization: multiple UEs transmit SRS, the base station captures the UAV-scattered echoes, a LOS-referenced synchronization that reuses only the existing timing-advance command and an adjacent-occasion conjugate product removes per-UE residuals to sub-nanosecond accuracy, and a multi-UE joint detector that accumulates normalized bistatic contrast over a shared 3D state space then localizes the UAV to 4.84 m median error in a cluttered urban FR1 simulation.

Load-bearing premise

That residual per-UE timing, frequency and amplitude errors can be cancelled to sub-nanosecond accuracy using only the existing timing-advance command and an adjacent-occasion conjugate product, without extra signaling, and that the resulting multi-UE fused contrast is then strong enough for reliable detection against LOS and urban clutter.

Editorial extensions

If this is right

Cellular base stations can turn ordinary phone SRS into a passive UAV radar without new spectrum, hardware or UE-side changes.
Multi-UE fusion over a shared 3D state space can lift weak bistatic UAV echoes above LOS and urban clutter that defeat single-UE detection.
Existing timing-advance commands become dual-use: they keep the air interface synchronized and simultaneously enable sub-nanosecond sensing coherence.
A 100 MHz FR1 uplink already supports roughly 5 m median 3D localization of low-altitude targets in cluttered cities.

Reading between the lines

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

The same LOS-referenced residual cancellation could be applied to other uplink pilots (e.g., DMRS) or to multi-static downlink sensing that also suffers from independent UE oscillators.
If the normalized bistatic contrast scales with UE density, denser urban deployments would tighten localization further without extra bandwidth.
The framework suggests a natural extension to joint communication-and-sensing resource allocation that schedules SRS occasions for both rate and sensing geometry.
Field validation with real multi-UE clocks and measured urban multipath would be the decisive next experiment.

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.

Desk Editor's Note

read the letter

The one thing to know is that this is an abstract-only pitch for the first uplink multi-UE SRS framework that does passive UAV localization at a 5G base station. They reuse existing TA commands plus an adjacent-occasion conjugate product for LOS-referenced sync, then run a fuse-then-detect joint search over a shared 3D state that accumulates normalized bistatic contrast across UEs. In one simulated cluttered urban FR1 scene with four pedestrian UEs and 100 MHz NR they report sub-nanosecond residual sync and 4.84 m median 3D error.

What is actually new is the uplink multi-UE angle itself. Most ISAC UAV work has been monostatic or downlink bistatic/multistatic; treating commodity UEs as opportunistic illuminators and cleaning their independent oscillators and timing loops without extra signaling is a real systems contribution if it holds. The two-stage design (sync first, then multi-UE accumulation) is sensible engineering and directly targets the two hard problems the abstract names: residual impairments that corrupt delay/Doppler, and an echo weaker than LOS plus urban clutter.

The soft spots are exactly what you would expect from abstract-only material. There are no equations, no ablation of the conjugate-product step, no baselines, no error bars, no false-alarm analysis, and no over-the-air data. We cannot tell whether the sub-ns claim survives realistic UE mobility, FR1 multipath, or residual frequency/amplitude offsets, nor whether the fused contrast actually keeps false alarms under control when the LOS is orders of magnitude stronger. The single geometry and four-UE setup leave open the usual questions about scaling and robustness. Those are not manufactured flaws; they are simply uninspectable from what we have.

This is for the ISAC and low-altitude sensing crowd who care about reusing commercial 5G infrastructure rather than adding radar. A serious referee should see the full paper; the problem is important enough and the approach is concrete enough that desk rejection would be premature. If the methods section delivers the missing verification, it is worth the community’s time. If not, the numbers stay unanchored. I would send it out.

Referee Report

3 major / 2 minor

Summary. The manuscript proposes what it describes as the first uplink passive UAV localization framework that exploits multi-UE 5G SRS pilots received at a base station. Residual per-UE timing, frequency, and amplitude impairments that corrupt UAV delay/Doppler are addressed by a LOS-referenced synchronization scheme that reuses the existing timing-advance command and an adjacent-occasion conjugate product, without extra signaling. A subsequent fuse-then-detect joint detector searches a shared 3D state space and accumulates multi-UE evidence via a normalized contrast that exploits bistatic geometry. Evaluation is reported for a single cluttered urban FR1 simulation with four pedestrian UEs and a 100 MHz NR waveform, claiming sub-nanosecond synchronization and a 4.84 m median 3D position error.

Significance. If the synchronization and multi-UE fusion claims hold under realistic impairments and clutter, the work would open a practically relevant ISAC path for low-altitude UAV sensing that uses commercial uplink SRS and existing TA signaling rather than dedicated radar or downlink-only configurations. Meter-level 3D localization from multi-UE uplink echoes would be of clear interest for airspace safety and privacy applications. The abstract’s emphasis on no additional signaling and a geometry-aware normalized contrast is a strength in principle; however, significance cannot be established from the abstract alone, because the load-bearing technical steps and the supporting evidence remain uninspectable.

major comments (3)

Abstract (synchronization claim): The central assertion that residual timing, frequency, and amplitude impairments are removed to sub-nanosecond accuracy solely by reusing the existing TA command plus an adjacent-occasion conjugate product is load-bearing for every subsequent detection and localization result. With only the abstract available, there are no equations, residual models, or validation curves that show cancellation under realistic UE oscillator offsets, timing-loop dynamics, mobility, and FR1 multipath. This claim must be derived and stress-tested explicitly before the reported sync accuracy can be accepted.
Abstract (joint-detector / performance claim): The reported 4.84 m median 3D position error is obtained from a single simulated urban scene with four UEs and 100 MHz NR. No error bars, no ablation of the sync stage versus the detector, no comparison baselines (single-UE, detect-then-fuse, conventional bistatic, or monostatic), and no real over-the-air measurements are visible. Without these, it is impossible to judge whether the multi-UE fused normalized contrast actually accumulates evidence without false-alarm inflation against stronger LOS and urban clutter, or whether the number is an artifact of the chosen geometry and simulation parameters.
Abstract (novelty / first-uplink claim): The paper asserts it is the first uplink multi-UE SRS passive UAV localization framework. That claim is material to significance and must be supported by a precise literature positioning against prior monostatic, downlink bistatic/multistatic, and any existing uplink ISAC or multi-static SRS sensing work. The abstract alone does not supply that positioning, so the novelty boundary remains unverified.

minor comments (2)

Abstract: Quantities such as “sub-nanosecond synchronization” and “4.84 m median 3D position error” should be accompanied, once the full text is available, by the precise metric definitions (e.g., residual delay RMSE after sync, CDF of 3D error) and by the number of Monte Carlo trials or geometry realizations used.
Abstract: The phrase “normalized contrast that exploits the bistatic geometry” is central to the detector but is left undefined at the abstract level; a short clarifying clause or pointer to the defining equation would help readers assess the fusion step.

Circularity Check

0 steps flagged · score 0.0 of 10

Abstract-only review: no demonstrable circularity; method-plus-simulation claims do not reduce to inputs by construction.

full rationale

Only the abstract is available, so no equations, derivation chain, self-citations, uniqueness theorems, or fitted-parameter steps can be inspected. The abstract states a design (LOS-referenced synchronization reusing TA and adjacent-occasion conjugate product; multi-UE joint detector with shared 3D search and bistatic normalized contrast) and reports simulation outcomes (sub-nanosecond synchronization; 4.84 m median 3D position error in a cluttered FR1 urban scene with four pedestrian UEs and 100 MHz NR). These are standard method-plus-evaluation claims. Nothing in the quoted text equates a predicted quantity to a fitted input by construction, renames a known empirical pattern as a first-principles result, or load-bears on an unverified self-citation. Residual risk that simulation geometry or parameters were tuned is not circularity under the stated rules, which require a quotable reduction. Score 0 with empty steps is therefore the honest finding.

Assumptions & free parameters 0 free parameters · 4 assumptions · 0 invented entities

Abstract-only: free parameters, simulation constants, and channel models are not enumerated. The claim rests on standard 5G NR uplink SRS and TA mechanisms, on the physical assumption that a UAV produces a usable bistatic echo weaker than LOS/clutter, and on the modeling assumption that residual impairments are removable by TA + adjacent-occasion conjugate product. No new physical entities are introduced.

assumptions (4)

domain assumption 5G NR uplink SRS pilots and existing timing-advance (TA) commands are available at the BS and can be reused for sensing without extra signaling.
Stated as the basis of the LOS-referenced synchronization scheme in the abstract.
domain assumption UAV-scattered echoes of multi-UE uplink SRS are present at the BS and, after residual removal, are detectable by multi-UE fusion against LOS and urban clutter.
Core sensing premise; abstract notes the echo is weaker than LOS and clutter, so detection from a single UE is unreliable.
ad hoc to paper Adjacent-occasion conjugate product plus TA removes residual timing, frequency, and amplitude impairments to sub-nanosecond accuracy.
Paper-specific synchronization construction; load-bearing for the reported sync and localization numbers.
ad hoc to paper A shared 3D state-space search with normalized contrast exploiting bistatic geometry correctly accumulates multi-UE evidence.
Paper-specific joint detector design claimed to enable reliable detection.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Fuse-then-Detect for Passive UAV Localization Using Multi-UE 5G Uplink Signals." pith.science (2026). https://pith.science/paper/RG4XAFBY

@misc{pith2026260711955,
  author       = {Pith},
  title        = {Pith review of: Fuse-then-Detect for Passive UAV Localization Using Multi-UE 5G Uplink Signals},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/RG4XAFBY}},
  note         = {Machine review of arXiv:2607.11955}
}

read the original abstract

Low-altitude uncrewed aerial vehicles (UAVs) can pose growing risks to airspace safety, security, and privacy. Cellular infrastructure can passively sense them without dedicated radar hardware by exploiting integrated sensing and communication (ISAC) technology. Most prior work exploits monostatic sensing or bistatic/multistatic configurations based on downlink measurements. To the best of our knowledge, this paper presents the first uplink framework, where multiple user equipments (UEs) transmit sounding reference signal (SRS) pilots and the base station (BS) receives the UAV-scattered echoes. Sensing from uplink SRS, however, introduces new challenges. Each UE has its own oscillator and timing loop, so the channel estimate at the BS carries residual timing, frequency, and amplitude impairments that corrupt the UAV delay and Doppler. Moreover, the UAV echo is weaker than both the line-of-sight (LOS) path and urban clutter, so detection from a single UE transmission is not reliable. We address these challenges by designing a LOS-referenced synchronization scheme and a joint detector. The synchronization reuses the existing timing advance (TA) command and an adjacent-occasion conjugate product to remove the residuals without additional signaling. Then the detector searches a shared 3D state space and accumulates evidence across UEs. It leverages a normalized contrast that exploits the bistatic geometry. We evaluate the framework in a cluttered urban scene at frequency range 1 (FR1) with four pedestrian UEs and a 100 MHz 5G New Radio (NR) waveform. The proposed pipeline achieves sub-nanosecond synchronization and a 4.84 m median 3D position error.

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 2 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

Multi-UAV Tracking Evaluation Using 5G Uplink Signals on an O-RAN ISAC Simulation Testbed
cs.NI 2026-08 conditional novelty 6.0 of 10

In an O-RAN simulation testbed, three UAVs are detected far more often than they are tracked because two targets often share one detection, and emulation wall-clock delays distort the tracking metrics.
Passive AoA Estimation of COTS 5G NR Handsets from Uplink SRS: A Practical USRP-B210 Implementation
eess.SP 2026-07 conditional novelty 6.0 of 10

Passive angle-of-arrival estimation of unmodified commercial 5G handsets from native uplink SRS is feasible with a two-element USRP B210 and a stock srsRAN gNB; accuracy is gated by SINR and multipath rather than range.

Pith tools

Reviewed July 15, 2026 · model on record in the stance chip above.