arxiv: 2602.17407 · v3 · submitted 2026-02-19 · 📡 eess.SY · cs.RO· cs.SY

Recognition: 2 theorem links

· Lean Theorem

Bluetooth Phased-array Aided Inertial Navigation Using Factor Graphs: Experimental Verification

Glen Hjelmerud M{\o}rkbak S{\o}rensen , Torleiv H. Bryne , Kristoffer Gryte , Tor Arne Johansen

Authors on Pith no claims yet

Pith reviewed 2026-05-15 21:09 UTC · model grok-4.3

classification 📡 eess.SY cs.ROcs.SY

keywords Bluetooth phased-arrayinertial navigationfactor graphsGNSS-deniedmultirotor dronerobust estimationaided navigation

0 comments

The pith

Bluetooth phased-array angular measurements enable usable factor-graph inertial navigation for drones during GNSS loss.

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

The paper tests low-cost commercial Bluetooth phased-array systems as an aid for inertial navigation on multirotor drones when GNSS is unavailable. It compares several robust estimation strategies inside a factor-graph optimizer against real flight data collected in loss-of-GNSS conditions. The results indicate that the noisy angular measurements, when combined with range or barometric pressure, can limit position error growth compared with unaided inertial navigation. A reader would care because the approach uses off-the-shelf parts rather than specialized radio hardware, lowering the cost for warehouse logistics, landings, and docking tasks.

Core claim

A factor graph optimisation-based estimator maintains usable navigation performance in GNSS-denied drone flights by incorporating Bluetooth phased-array angular measurements together with range or barometric pressure data. The work evaluates multiple robust estimation strategies on experimental multirotor flight records and shows that the Bluetooth angular factors remain sufficiently informative despite elevated noise levels and short range.

What carries the argument

Factor graph optimisation-based estimator that adds Bluetooth phased-array angular measurements as measurement factors alongside inertial, range, and pressure terms.

If this is right

Navigation error remains bounded rather than growing without bound during GNSS outages when Bluetooth angular measurements are fused in the factor graph.
Robust estimation strategies inside the factor graph reduce the impact of the higher noise present in commercial Bluetooth phased-array readings.
Adding barometric pressure or range measurements alongside the Bluetooth angles improves vertical and horizontal accuracy in the same estimator.
The same factor-graph structure supports low-cost COTS hardware for repeated drone operations in GNSS-denied indoor or warehouse environments.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

The same Bluetooth-aided factor graph could be tested on ground robots or fixed-wing vehicles that already carry Bluetooth infrastructure.
Improving the statistical model of the phased-array angle noise would likely extend the usable range beyond what the current experiments demonstrate.
The approach suggests that other short-range radio angle sensors could be swapped into the factor graph with only modest retuning of the robust cost functions.

Load-bearing premise

The Bluetooth phased-array angular measurements stay informative enough and can be modeled accurately enough that the factor-graph estimator produces usable position estimates even though the signals are noisier and shorter-range than dedicated systems.

What would settle it

If the position and attitude errors in the GNSS-denied flight segments grow at the same rate as unaided inertial navigation or diverge when the Bluetooth angular factors are included, the claim of usable aided performance would be falsified.

read the original abstract

Phased-array Bluetooth systems have emerged as a low-cost alternative for performing aided inertial navigation in GNSS-denied use cases such as warehouse logistics, drone landings, and autonomous docking. Basing a navigation system off of commercial-off-the-shelf components may reduce the barrier of entry for phased-array radio navigation systems, albeit at the cost of significantly noisier measurements and relatively short feasible range. In this paper, we compare robust estimation strategies for a factor graph optimisation-based estimator using experimental data collected from multirotor drone flight. We evaluate performance in loss-of-GNSS scenarios when aided by Bluetooth angular measurements, as well as range or barometric pressure.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

Bluetooth phased-array angles add some help to factor-graph inertial navigation on real drone flights but the gains look modest and the paper needs hard numbers on error reduction to show they matter.

read the letter

The main point is that this paper runs real multirotor drone flights and folds commercial Bluetooth phased-array angular measurements into a factor-graph inertial estimator for GNSS-denied segments, testing them together with range or barometric factors and comparing robust strategies on the collected data. It shows the setup can run on low-cost hardware for uses like warehouse logistics or indoor operations. The experimental data collection and the direct comparison of estimation approaches on actual loss-of-GNSS segments are the useful parts. Factor graphs fit naturally here since they let the authors add the radio measurements as extra factors without rebuilding the whole estimator. That practical angle is worth noting for people who need to work with off-the-shelf sensors rather than specialized phased arrays. The soft spots sit in the evidence for how much the Bluetooth terms actually help. The measurements are noisier and shorter-range than dedicated systems, yet the description gives no position or velocity error numbers, no drift reduction figures, and no breakdown of how much the angular factors cut uncertainty compared with IMU-plus-baro or IMU-plus-range baselines. Without those metrics or details on the bearing model and any robust weighting, it is difficult to tell whether the Bluetooth data drives real improvement or mostly rides along while the other sensors do the work. The single-platform flights also leave open how the results hold up in different multipath environments. This is the sort of incremental experimental paper aimed at navigation engineers building aided inertial systems for small UAVs or robots. Readers who want to see factor-graph methods applied to real radio data will get some value from the comparison, even if the Bluetooth contribution turns out limited. It deserves peer review so referees can ask for the missing quantitative results and check the measurement modeling.

Referee Report

3 major / 2 minor

Summary. The manuscript presents an experimental comparison of robust estimation strategies within a factor-graph optimization framework for inertial navigation of a multirotor drone. It evaluates the addition of Bluetooth phased-array angular measurements, combined with either range or barometric pressure factors, during GNSS-denied flight segments using collected real-world data.

Significance. If the results demonstrate that the Bluetooth angular factors produce measurable reduction in drift relative to IMU-only or IMU+baro baselines despite higher noise and shorter range, the work would support low-cost COTS phased-array aiding for GNSS-denied navigation in logistics and docking scenarios. The factor-graph formulation is a natural fit for incorporating heterogeneous, noisy measurements, and the experimental focus on real flight data is a strength.

major comments (3)

[Abstract] Abstract: the description of the experimental comparison supplies no quantitative error metrics (e.g., RMSE or drift rates), data exclusion rules, or explicit definition of the robust strategies, preventing verification that the collected measurements support the performance claims.
[Results / Experimental Evaluation] The central claim that Bluetooth angular measurements remain sufficiently informative rests on an untested premise; the manuscript should isolate their contribution by reporting separate runs with and without the Bluetooth factors (while keeping range/baro fixed) to quantify information gain versus the stronger barometric or range terms.
[Measurement Model / Data Collection] No details are given on the effective range, SNR variation, or multipath-induced bias observed in the Bluetooth measurements during the flights; without these, it is impossible to assess whether the measurement model (presumably a bearing factor with additive Gaussian noise) produces usable information or merely adds inconsistency.

minor comments (2)

[Factor Graph Formulation] Notation for the factor-graph variables and the robust cost functions should be defined explicitly in a dedicated section or table rather than introduced inline.
[Figures] Figure captions should state the exact flight segments, sensor configurations, and any data filtering applied so that readers can reproduce the plotted trajectories.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for the constructive and detailed feedback on our manuscript. We have revised the paper to strengthen the abstract with quantitative metrics, add explicit ablation studies isolating the Bluetooth angular factors, and include characterization of the Bluetooth measurement statistics. Point-by-point responses follow.

read point-by-point responses

Referee: [Abstract] Abstract: the description of the experimental comparison supplies no quantitative error metrics (e.g., RMSE or drift rates), data exclusion rules, or explicit definition of the robust strategies, preventing verification that the collected measurements support the performance claims.

Authors: We agree that the abstract should convey the key quantitative outcomes. In the revised version we have expanded the abstract to report RMSE values for position and attitude during the GNSS-denied segments, stated the data exclusion criteria (SNR threshold and 3-sigma residual test), and named the robust estimators (Huber loss with fixed tuning constants). These additions directly address the concern while preserving the abstract's brevity. revision: yes
Referee: [Results / Experimental Evaluation] The central claim that Bluetooth angular measurements remain sufficiently informative rests on an untested premise; the manuscript should isolate their contribution by reporting separate runs with and without the Bluetooth factors (while keeping range/baro fixed) to quantify information gain versus the stronger barometric or range terms.

Authors: We accept the suggestion to make the incremental value explicit. The revised results section now contains dedicated ablation tables that compare IMU+baro versus IMU+baro+Bluetooth and IMU+range versus IMU+range+Bluetooth on the same flight segments. The tables report drift rates and final position errors, demonstrating a measurable reduction attributable to the angular factors even when the stronger barometric or range terms are already present. revision: yes
Referee: [Measurement Model / Data Collection] No details are given on the effective range, SNR variation, or multipath-induced bias observed in the Bluetooth measurements during the flights; without these, it is impossible to assess whether the measurement model (presumably a bearing factor with additive Gaussian noise) produces usable information or merely adds inconsistency.

Authors: We acknowledge the omission. A new subsection has been added to the experimental setup that reports the observed effective range (typically 10–35 m), SNR statistics across the flights (median 18 dB, range 8–28 dB), and a brief multipath analysis (occasional 3–6° biases identified via residual inspection). We confirm the factor is a bearing measurement with additive Gaussian noise and include a residual histogram to support model consistency. revision: yes

Circularity Check

0 steps flagged

No derivation chain present; purely experimental comparison

full rationale

The paper describes an experimental verification of factor-graph estimators on collected multirotor flight data in GNSS-denied scenarios, comparing Bluetooth angular measurements against range and barometric aids. No new equations, parameter fittings, uniqueness theorems, or ansatzes are introduced that could reduce to self-definition or self-citation. All performance claims rest on direct comparison of optimizer outputs to ground-truth trajectories from the flight logs, with no load-bearing mathematical steps that loop back to the inputs by construction.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Abstract supplies no information on free parameters, axioms, or invented entities; full manuscript would be required to audit these.

pith-pipeline@v0.9.0 · 5431 in / 967 out tokens · 74737 ms · 2026-05-15T21:09:47.228927+00:00 · methodology

discussion (0)

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

IndisputableMonolith/Cost/FunctionalEquation.lean washburn_uniqueness_aczel unclear

?

unclear
Relation between the paper passage and the cited Recognition theorem.

The estimation itself is based on minimising ˆT=argminT∥h(T)−z∥2P ... on-manifold IMU preintegration factor ... Tukey and Geman-McClure M-estimators
IndisputableMonolith/Foundation/RealityFromDistinction.lean reality_from_one_distinction unclear

?

unclear
Relation between the paper passage and the cited Recognition theorem.

SE(3) matrix Lie group ... Bluetooth LE Direction Finding ... handover scenarios

What do these tags mean?

matches: The paper's claim is directly supported by a theorem in the formal canon.
supports: The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends: The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses: The paper appears to rely on the theorem as machinery.
contradicts: The paper's claim conflicts with a theorem or certificate in the canon.
unclear: Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.