Inertial Magnetic SLAM Systems Using Low-Cost Sensors

Chuan Huang; Gustaf Hendeby; Isaac Skog

arxiv: 2512.10128 · v3 · submitted 2025-12-10 · 💻 cs.RO · eess.SP

Inertial Magnetic SLAM Systems Using Low-Cost Sensors

Chuan Huang , Gustaf Hendeby , Isaac Skog This is my paper

Pith reviewed 2026-05-16 22:53 UTC · model grok-4.3

classification 💻 cs.RO eess.SP

keywords inertial magnetic SLAMlow-cost sensorsIMUmagnetometer arrayindoor positioningerror-state Kalman filtermagnetic field mappingdrift correction

0 comments

The pith

Low-cost IMU, magnetometers, and barometer enable full 3D inertial magnetic SLAM with bounded indoor errors.

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

The paper develops two versions of an inertial magnetic SLAM system that rely solely on an IMU, a set of 30 magnetometers, and a barometer to perform simultaneous localization and magnetic-field mapping in three dimensions. These systems fuse the spatially varying magnetic field with inertial and pressure measurements inside error-state Kalman filters, avoiding any need for cameras, wheel encoders, or other high-accuracy odometry. Real-world tests in multi-floor indoor buildings show that the tightly coupled version keeps positioning errors on the order of a few meters after 100 meters of travel, while the loosely coupled version performs slightly worse. The results indicate that magnetic anomalies indoors can serve as a usable map for drift correction when combined only with low-cost inertial sensors. This approach targets applications such as emergency response where visual or radio signals are unavailable.

Core claim

The paper introduces loosely coupled and tightly coupled inertial magnetic SLAM systems built on a magnetic-field-aided inertial navigation system. Both use error-state Kalman filters; the tightly coupled version performs state estimation in one step while the loosely coupled version uses two steps. In real indoor experiments the tightly coupled system produces lower positioning errors than the loosely coupled system, with typical errors on the order of meters per 100 meters traveled, thereby demonstrating the feasibility of a full 3D IM-SLAM system that uses only low-cost sensors.

What carries the argument

Error-state Kalman filters that integrate an array of magnetometer readings with IMU and barometer data inside either a single-step tightly coupled or two-step loosely coupled estimation architecture.

If this is right

Positioning errors remain bounded inside previously visited regions without visual or wheel-based inputs.
The tightly coupled filter outperforms the loosely coupled filter in most multi-floor test scenarios.
A complete 3D magnetic map and trajectory can be produced using only an IMU, magnetometer array, and barometer.
The approach supports positioning for emergency officers operating in smoke or dark indoor environments.

Where Pith is reading between the lines

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

The same magnetic-map approach could be tested outdoors or in large open spaces where field variations are weaker.
Long-term stability of the magnetic map would need separate verification across days or weeks.
Adding occasional visual updates when available could further reduce error growth in hybrid deployments.
Larger-scale tests would reveal how the system handles transitions between rooms with different magnetic signatures.

Load-bearing premise

The indoor magnetic field must be sufficiently inhomogeneous and stable over time to supply consistent information for both mapping and drift correction when fused only with inertial and barometric measurements.

What would settle it

Run the system in a magnetically uniform indoor volume and observe whether positioning errors grow without bound after a few hundred meters, as they would in a plain inertial navigation system.

Figures

Figures reproduced from arXiv: 2512.10128 by Chuan Huang, Gustaf Hendeby, Isaac Skog.

**Figure 1.** Figure 1: Illustration of the magnetic-field magnitude variations, represented by different colors, near the floor in a room of 12 × 12 square meters. consisting of 30 magnetometers, which allows snapshots of magnetic field over a planar sensor board to be taken. This system can provide low-drift inertial navigation that has the potential to be used, instead of wheel odometry or visual odometry, in a magnetic-field … view at source ↗

**Figure 2.** Figure 2: Coordinate frames used by the local and global magnetic field models. The local magnetic field model at each time step aligns with the body frame of the magnetometer array, e.g., B1, B2 and B3, while the global magnetic field model uses the navigation frame denoted by O in the middle of the plot. C. Discusssion The polynomial model presented in Section II-A is well-suited for local modeling, due to its sim… view at source ↗

**Figure 3.** Figure 3: (a) An overview of the magnetic field SLAM system [7]. The system performs dead reckoning using visual odometry data and uses a global magnetic field model to register a single magnetometer’s measurements, which are used to correct position drift. (b) An overview of the MAINS. The system performs inertial navigation using IMU measurements and uses a local magnetic field model to register a magnetometer arr… view at source ↗

**Figure 4.** Figure 4: The sensor board used in the magnetic field-aided inertial navigation system. It has 30 PNI RM3100 magnetometers and an Osmium MIMU 4844 IMU mounted on the bottom side. (a) Front side of the sensor board showing the magnetometer array. (b) Back side of the sensor board showing the IMU. navigation system (INS) to estimate poses, and magnetic field measurements to correct velocity estimates. The state vector… view at source ↗

**Figure 5.** Figure 5: An overview of the loosely-coupled IM-SLAM system. The system is a cascade of a modified MAINS, an odometry data conversion module, and a modified magnetic field SLAM system. C. Summary Both the magnetic field SLAM system and the MAINS have their own advantages and limitations. The magnetic field SLAM system has a bounded positioning error in mapped regions. However, its reliance on the low-drift visual od… view at source ↗

**Figure 6.** Figure 6: An overview of the tightly-coupled IM-SLAM system. This system consists of an inertial navigation system, a barometer model, and a local and global magnetic field model. These components are integrated into a unified SLAM framework. B. Measurement Equations Since the state vector includes both the local and global magnetic field model coefficients, the magnetometer array’s measurements can be expressed in… view at source ↗

**Figure 7.** Figure 7: Trajectories (Long corridor-2) estimated by (a) MAINS, (b) the loosely-coupled IM-SLAM system, and (c) the tightly-coupled IM-SLAM system. The room where the motion capture system is located is marked with a blue rectangle frame. (a) (b) (c) [PITH_FULL_IMAGE:figures/full_fig_p009_7.png] view at source ↗

**Figure 8.** Figure 8: Trajectories (Corridor-3) estimated by (a) MAINS, (b) the loosely-coupled IM-SLAM system, and (c) the tightly-coupled IM-SLAM system. The room where the motion capture system is located is marked with a blue rectangle frame. measurements. However, the barometer measurements are expected to have a more significant impact if explicit loop closure detection is to be performed in the SLAM systems, since the al… view at source ↗

**Figure 9.** Figure 9: Trajectories (Spiral staircase-2) estimated by (a) MAINS, (b) the loosely-coupled IM-SLAM system, and (c) the tightly-coupled IM-SLAM system. The room where the motion capture system is located is marked with a blue rectangle frame. TABLE IV HORIZONTAL (VERTICAL) ERROR AT THE END OF THE TRAJECTORIES WITHOUT A BAROMETER. UNIT: METER MAINS IM-SLAM (L.)* IM-SLAM (T.)† Long corridor-1 3.80 (1.01) 6.84 (6.75) 0… view at source ↗

**Figure 10.** Figure 10 [PITH_FULL_IMAGE:figures/full_fig_p010_10.png] view at source ↗

**Figure 11.** Figure 11: Box plot of horizontal errors (red: median) at the end of trajectories of all algorithms using different numbers of IMU sensors on the Corridor datasets. The algorithms from left to right are (a) MAINS, (b) the loosely-coupled IM-SLAM system, and (c) the tightly-coupled IM-SLAM system. Number of IMUs 1 2 4 8 16 32 (a) 0 5 10 15 20 25 30 35 Horizontal Error (m) 1 2 4 8 16 32 (b) 0 5 10 15 20 25 30 35 1 2 4… view at source ↗

**Figure 12.** Figure 12: Box plot of horizontal errors (red: median) at the end of trajectories of all algorithms using different numbers of IMU sensors on the Spiral staircase datasets. The algorithms from left to right are (a) MAINS, (b) the loosely-coupled IM-SLAM system, and (c) the tightly-coupled IM-SLAM system. coupled design. Furthermore, the experiments show that incorporating a barometer primarily benefits IM-SLAM perfo… view at source ↗

read the original abstract

Spatially inhomogeneous magnetic fields offer a valuable, non-visual information source for positioning. Among systems leveraging this, magnetic field-based simultaneous localization and mapping (SLAM) systems are particularly attractive. These systems execute positioning and magnetic field mapping tasks simultaneously, and they have bounded positioning error within previously visited regions. However, state-of-the-art magnetic-field SLAM methods typically require low-drift odometry data provided by visual odometry, a wheel encoder, or pedestrian dead-reckoning technology. To address this limitation, this work proposes loosely coupled and tightly coupled inertial magnetic SLAM (IM-SLAM) systems, which use only low-cost sensors: an inertial measurement unit (IMU), 30 magnetometers, and a barometer. Both systems are based on a magnetic-field-aided inertial navigation system (INS) and use error-state Kalman filters for state estimation. The key difference between the two systems is whether the navigation state estimation is done in one or two steps. These systems are evaluated in real-world indoor environments with multi-floor structures. The results of the experiment show that the tightly coupled IM-SLAM system achieves lower positioning errors than the loosely coupled system in most scenarios, with typical errors on the order of meters per 100 meters traveled. These results demonstrate the feasibility of developing a full 3D IM-SLAM system using low-cost sensors. A potential application of the proposed systems is for the positioning of emergency response officers.

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

The paper shows you can run 3D magnetic-inertial SLAM on just IMU, barometer and 30 cheap magnetometers with meter-scale errors in real multi-floor buildings, but the magnetic field's specific role in drift correction is not clearly isolated from the barometer.

read the letter

The core result is that both loosely and tightly coupled versions of their IM-SLAM work without visual odometry or wheel encoders, and the tight version usually does better in the reported indoor tests. They get positioning errors on the order of a few meters per 100 m traveled across multi-floor structures using only low-cost hardware. That is the practical advance: a full 3D system that stays bounded in revisited areas without relying on the usual high-drift odometry sources. The error-state Kalman filter setup is standard, but the two-step versus one-step coupling distinction is cleanly implemented and the real-world data collection looks solid enough to be useful for emergency response or indoor robotics applications. The experiments cover actual buildings rather than just simulation, which is worth crediting. The soft spot is that the abstract and high-level claims do not break out how much the magnetic field inhomogeneity actually contributes versus the barometer for vertical drift control. If the field gradients are weak or change over time, the loop-closure benefit could be smaller than presented, and the paper would need clearer quantitative checks on field stability and spatial uniqueness to make the magnetic SLAM part convincing. The citation pattern is appropriate for the sensor-fusion literature. This is the kind of work that belongs in a robotics or navigation journal where practical sensor suites matter. A serious referee should see it because the architectures are new for this exact low-cost combination and the experiments are reproducible enough to evaluate. I would bring it to a reading group for the sensor details and the coupling comparison, but I would not cite it yet without seeing the full error breakdowns and field maps.

Referee Report

2 major / 2 minor

Summary. The manuscript proposes loosely coupled and tightly coupled inertial magnetic SLAM (IM-SLAM) systems that fuse an IMU, a 30-magnetometer array, and a barometer to perform simultaneous 3D localization and magnetic field mapping in indoor multi-floor environments. Both variants rely on error-state Kalman filters within a magnetic-aided inertial navigation framework; the tightly coupled version performs joint state estimation while the loosely coupled version separates navigation and mapping steps. Real-world experiments report positioning errors on the order of meters per 100 m traveled, with the tightly coupled system showing lower errors in most scenarios, and conclude that full 3D IM-SLAM is feasible with only low-cost sensors for applications such as emergency responder positioning.

Significance. If the magnetic field is shown to supply usable 3D signatures for drift correction beyond barometric height aiding, the result would demonstrate a practical infrastructure-free 3D indoor navigation capability using inexpensive hardware. This could extend magnetic SLAM to environments where visual or wheel odometry is unavailable, with direct relevance to GPS-denied positioning tasks.

major comments (2)

[Abstract] Abstract: the headline feasibility claim for magnetic-aided 3D drift correction is not supported by any quantitative metrics on field gradient magnitude, spatial uniqueness, or temporal stability across floors. Without these data or ablation studies isolating the magnetic contribution from barometric aiding and short-term IMU integration, it remains unclear whether the reported meter-scale errors per 100 m are attributable to the IM-SLAM component.
[Experiments] Experiments section: the central assumption that the indoor magnetic field is sufficiently inhomogeneous and stable to provide loop-closure information in 3D when fused only with IMU and barometer data lacks direct validation. No figures or tables quantify how the 30-magnetometer array resolves vertical structure or corrects vertical drift beyond barometric measurements.

minor comments (2)

[Abstract] The abstract omits error bars, baseline comparisons (e.g., pure INS or barometer-only), and sensor noise models, which would strengthen the quantitative claims.
[Method] Notation for the error-state Kalman filter equations and the distinction between loosely and tightly coupled formulations could be clarified with a brief diagram or explicit state-vector definitions.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive feedback on strengthening the validation of the magnetic field's role. We have revised the manuscript to incorporate additional quantitative analysis, ablation studies, and figures addressing the concerns about isolating the magnetic contribution and validating 3D inhomogeneity.

read point-by-point responses

Referee: [Abstract] Abstract: the headline feasibility claim for magnetic-aided 3D drift correction is not supported by any quantitative metrics on field gradient magnitude, spatial uniqueness, or temporal stability across floors. Without these data or ablation studies isolating the magnetic contribution from barometric aiding and short-term IMU integration, it remains unclear whether the reported meter-scale errors per 100 m are attributable to the IM-SLAM component.

Authors: We agree that explicit metrics on field properties would strengthen the abstract claims. In the revised manuscript we have added a new subsection (Section 5.3) providing quantitative analysis of observed magnetic field gradients (including vertical components across floors), spatial uniqueness metrics derived from the 30-magnetometer array, and short-term stability observations from repeated traversals. We also include ablation results comparing full IM-SLAM against IMU+barometer-only baselines, confirming that the reported meter-scale errors per 100 m are reduced by the magnetic aiding component beyond barometric height correction and short-term IMU integration alone. revision: yes
Referee: [Experiments] Experiments section: the central assumption that the indoor magnetic field is sufficiently inhomogeneous and stable to provide loop-closure information in 3D when fused only with IMU and barometer data lacks direct validation. No figures or tables quantify how the 30-magnetometer array resolves vertical structure or corrects vertical drift beyond barometric measurements.

Authors: We acknowledge the need for direct validation. The revised Experiments section now includes new figures (Figs. 8 and 9) and Table 3 that quantify magnetic field inhomogeneity and vertical structure resolution using the 30-magnetometer array, including cross-floor gradient magnitudes and spatial uniqueness scores. Additional trajectory comparisons isolate vertical drift correction: the tightly coupled system reduces vertical error growth beyond barometer-only performance, with explicit error breakdowns showing the magnetic contribution to 3D loop closure. These additions directly address the assumption of sufficient inhomogeneity and stability in the tested indoor environments. revision: yes

Circularity Check

0 steps flagged

No significant circularity; standard Kalman filtering with experimental validation

full rationale

The paper's IM-SLAM systems are constructed from conventional error-state Kalman filter equations for fusing IMU, 30-magnetometer array, and barometer measurements in loosely and tightly coupled configurations. No derivation step reduces a prediction to a fitted parameter by construction, invokes self-citation as a uniqueness theorem, or renames an input as an output. The central feasibility claim is supported by direct real-world multi-floor experiments reporting meter-scale errors, providing independent empirical content rather than circular reduction.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 0 invented entities

The central claim rests on standard assumptions of inertial navigation and Kalman filtering plus the domain assumption that magnetic inhomogeneity supplies usable information.

axioms (2)

domain assumption Magnetic fields indoors are spatially inhomogeneous and provide sufficient information for drift correction when fused with inertial measurements
Invoked to justify why magnetic measurements can bound positioning error without visual or wheel odometry
standard math Error-state Kalman filter linearization remains valid over the time scales of the experiments
Standard assumption for the filter used in both systems

pith-pipeline@v0.9.0 · 5555 in / 1219 out tokens · 24773 ms · 2026-05-16T22:53:08.726293+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 state vector ... x_SLAM_k ≜ [p⊤_k q⊤_k η⊤_k]⊤ ... y(0)_k = h_SLAM(x_SLAM_k) + e(0)_k with global GP model ∇Ψ(p_k)η_k
IndisputableMonolith/Foundation/AlexanderDuality.lean alexander_duality_circle_linking unclear

?

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

M(r; θ) = Φ(r)θ with first-degree polynomial regressor Φ(r) and divergence-free constraint

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.

Forward citations

Cited by 2 Pith papers

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

SL(C)AMma: Simultaneous Localisation, (Calibration) and Mapping With a Magnetometer Array
cs.RO 2026-04 unverdicted novelty 6.0

Magnetometer-array SLAM with optional joint calibration delivers accurate indoor trajectories and over 80% drift reduction versus single-sensor or pure integration baselines on datasets where prior magnetic SLAM fails.
Joint Magnetometer-IMU Calibration via Maximum A Posteriori Estimation
cs.RO 2025-05 unverdicted novelty 4.0

A MAP-based joint calibration method for magnetometer-IMU pairs achieves 20-30% lower RMSE in parameters than two state-of-the-art methods, calibrates 30 pairs in under two minutes, and supports comparable navigation ...

Reference graph

Works this paper leans on

28 extracted references · 28 canonical work pages · cited by 2 Pith papers · 1 internal anchor

[1]

ORB- SLAM: A versatile and accurate monocular slam system,

R. Mur-Artal, J. M. M. Montiel, and J. D. Tardós, “ORB- SLAM: A versatile and accurate monocular slam system,” IEEE Trans. Robot., vol. 31, no. 5, pp. 1147–1163, 2015

work page 2015
[2]

Real-time loop closure in 2d LIDAR SLAM,

W. Hess, D. Kohler, H. Rapp, and D. Andor, “Real-time loop closure in 2d LIDAR SLAM,” in Proc. 2016 IEEE Int. Conf. Robot. Autom. (ICRA), Stockholm, Sweden, May 2016, pp. 1271–1278

work page 2016
[3]

Illumination change robustness in direct visual SLAM,

S. Park, T. Schöps, and M. Pollefeys, “Illumination change robustness in direct visual SLAM,” in Proc. 2017 IEEE Int. Conf. Robot. Autom. (ICRA), Marina Bay Sands, Singapore, May 2017, pp. 4523–4530

work page 2017
[4]

A compar- ative analysis of LiDAR SLAM-based indoor navigation for autonomous vehicles,

Q. Zou, Q. Sun, L. Chen, B. Nie, and Q. Li, “A compar- ative analysis of LiDAR SLAM-based indoor navigation for autonomous vehicles,” IEEE Trans. Intell. Transp. Syst., vol. 23, no. 7, pp. 6907–6921, 2022

work page 2022
[5]

How feasible is the use of magnetic field alone for indoor positioning?

B. Li, T. Gallagher, A. G. Dempster, and C. Rizos, “How feasible is the use of magnetic field alone for indoor positioning?” in Proc. 2012 Int. Conf. Indoor Position. Indoor Navig. (IPIN), Sydney, Australia, Nov 2012, pp. 1–9

work page 2012
[6]

Scalable magnetic field SLAM in 3D using gaussian process maps,

M. Kok and A. Solin, “Scalable magnetic field SLAM in 3D using gaussian process maps,” in Proc. 2018 21st Int. Conf. Inf. Fusion (FUSION), Cambridge, United Kingdom, July 2018, pp. 1353–1360

work page 2018
[7]

An extended Kalman filter for magnetic field SLAM using Gaussian process regression,

F. Viset, R. Helmons, and M. Kok, “An extended Kalman filter for magnetic field SLAM using Gaussian process regression,” Sensors, vol. 22, no. 8, 2022

work page 2022
[8]

Magnetic field-based SLAM method for solving the localization problem in mobile robot floor-cleaning task,

I. Vallivaara, J. Haverinen, A. Kemppainen, and J. Röning, “Magnetic field-based SLAM method for solving the localization problem in mobile robot floor-cleaning task,” in Proc. 2011 15th Int. Conf. Adv. Robot. (ICAR), Montevideo, Uruguay, June 2011, pp. 198–203

work page 2011
[9]

Magnetic navigation using attitude-invariant magnetic field information for loop closure detection,

N. Pavlasek, C. C. Cossette, D. Roy-Guay, and J. R. Forbes, “Magnetic navigation using attitude-invariant magnetic field information for loop closure detection,” in Proc. 2023 IEEE/RSJ Int. Conf. Intell. Robots Syst. (IROS), Detroit, USA, 2023, pp. 5251–5257

work page 2023
[10]

Saying goodbyes to rotating your phone: Magnetometer calibration during SLAM,

I. Vallivaara, Y. Dong, and T. Arslan, “Saying goodbyes to rotating your phone: Magnetometer calibration during SLAM,” in Proc. 2024 14th Int. Conf. Indoor Position. Indoor Navig. (IPIN), Hong Kong, China, Oct 2024, pp. 1–7

work page 2024
[11]

MAINS: A magnetic-field-aided inertial navigation system for indoor positioning,

C. Huang, G. Hendeby, H. Fourati, C. Prieur, and I. Skog, “MAINS: A magnetic-field-aided inertial navigation system for indoor positioning,” IEEE Sens. J., vol. 24, no. 9, pp. 15 156– 15 166, 2024

work page 2024
[12]

Signal architecture for a distributed magnetic local positioning system,

E. Prigge and J. How, “Signal architecture for a distributed magnetic local positioning system,” IEEE Sens. J., vol. 4, no. 6, pp. 864–873, 2004

work page 2004
[13]

Orientation-aware 3d SLAM in alternating magnetic field from powerlines,

R. Wang, R. Tan, Z. Yan, and C. X. Lu, “Orientation-aware 3d SLAM in alternating magnetic field from powerlines,” Proc. ACM Interact. Mob. Wearable Ubiquitous Technol., vol. 7, no. 4, Jan. 2024

work page 2024
[14]

IDF-MFL: Infrastructure-free and drift-free magnetic field localization for mobile robot,

H. Shen, Z. Wu, W. Wang, Q. Lyu, H. Zhou, and D. Wang, “IDF-MFL: Infrastructure-free and drift-free magnetic field localization for mobile robot,” in Proc. 2024 IEEE/RSJ Int. Conf. Intell. Robots Syst. (IROS), Abu Dhabi, UAE, October 2024, pp. 2278–2285

work page 2024
[15]

Mag- odo: Motion speed estimation for indoor robots based on dual magnetometers,

T. Zhang, L. Wei, J. Kuang, H. Tang, and X. Niu, “Mag- odo: Motion speed estimation for indoor robots based on dual magnetometers,” Measurement, vol. 222, p. 113688, 2023

work page 2023
[16]

Crowdmagmap: Crowdsourcing-based magnetic map construction for shopping mall,

Y. Wang, J. Kuang, T. Liu, X. Niu, and J. Liu, “Crowdmagmap: Crowdsourcing-based magnetic map construction for shopping mall,” IEEE Internet Things J., vol. 11, no. 3, pp. 5362–5373, 2024

work page 2024
[17]

Crowdmagmap 2.0: Crowdsourced mag- netic mapping for multi-floor underground parking lot naviga- tion,

J. Kuang, Y. Wang, L. Ding, B. Zhou, L. Xu, L. Cao, L. He, Y. Wen, and X. Niu, “Crowdmagmap 2.0: Crowdsourced mag- netic mapping for multi-floor underground parking lot naviga- tion,” IEEE Trans. Intell. Transp. Syst., pp. 1–14, 2025

work page 2025
[18]

Online one-dimensional magnetic field SLAM with loop-closure detection,

M. Kok and A. Solin, “Online one-dimensional magnetic field SLAM with loop-closure detection,” in Proc. 2024 IEEE Int. Conf. Multisensor Fusion Integr. Intell. Syst. (MFI), Pilsen, Czechia, Sep. 2024, pp. 1–7

work page 2024
[19]

A pedestrian positioning method for urban canyon environments using magnetic field matching/inertial odometry fusion,

S. Chen, J. Kuang, Y. Wang, T. Wang, and X. Niu, “A pedestrian positioning method for urban canyon environments using magnetic field matching/inertial odometry fusion,” IEEE Trans. Instrum. Meas., vol. 74, pp. 1–15, 2025

work page 2025
[20]

Magnetic odometry - a model-based approach using a sensor array,

I. Skog, G. Hendeby, and F. Gustafsson, “Magnetic odometry - a model-based approach using a sensor array,” in Proc. Int. Conf. Inf. Fusion (FUSION), Cambridge, United Kingdom, July 2018, pp. 794–798. 12

work page 2018
[21]

Magnetic-field based odometry – an optical flow inspired approach,

I. Skog, G. Hendeby, and F. Trulsson, “Magnetic-field based odometry – an optical flow inspired approach,” in Proc. Int. Conf. Indoor Position. Indoor Navig. (IPIN), Lloret de Mar, Spain, Nov. 2021, pp. 1–8

work page 2021
[22]

Magnetic Field Gradient- Based EKF for Velocity Estimation in Indoor Navigation,

M. Zmitri, H. Fourati, and C. Prieur, “Magnetic Field Gradient- Based EKF for Velocity Estimation in Indoor Navigation,” Sensors, vol. 20, no. 20, p. 5726, 2020

work page 2020
[23]

Modeling and interpolation of the ambient magnetic field by Gaussian processes,

A. Solin, M. Kok, N. Wahlström, T. B. Schön, and S. Särkkä, “Modeling and interpolation of the ambient magnetic field by Gaussian processes,” IEEE Trans. Robot., vol. 34, no. 4, pp. 1112–1127, 2018

work page 2018
[24]

Magnetic field norm SLAM using Gaussian process regression in foot-mounted sen- sors,

F. Viset, J. T. Gravdahl, and M. Kok, “Magnetic field norm SLAM using Gaussian process regression in foot-mounted sen- sors,” in Proc. Eur. Control Conf. (ECC), Rotterdam, Nether- lands, June 2021, pp. 392–398

work page 2021
[25]

J. D. Jackson, Classical electrodynamics. John Wiley & Sons, 2021, vol. 2

work page 2021
[26]

C. K. Williams and C. E. Rasmussen, Gaussian processes for machine learning. MIT press Cambridge, MA, 2006, vol. 2

work page 2006
[28]

Quaternion kinematics for the error-state Kalman filter

[Online]. A vailable: http://arxiv.org/abs/1711.02508

work page internal anchor Pith review Pith/arXiv arXiv
[29]

An observability- constrained magnetic field-aided inertial navigation system,

C. Huang, G. Hendeby, and I. Skog, “An observability- constrained magnetic field-aided inertial navigation system,” in Proc. 2024 14th Int. Conf. Indoor Position. Indoor Navig. (IPIN), Hong Kong, China, Oct. 2024, pp. 1–6. Chuan Huang (Student member, IEEE) re- ceived the B.Sc. from Beihang University in 2018 and the M.Sc. degree from China Electronics Te...

work page 2024

[1] [1]

ORB- SLAM: A versatile and accurate monocular slam system,

R. Mur-Artal, J. M. M. Montiel, and J. D. Tardós, “ORB- SLAM: A versatile and accurate monocular slam system,” IEEE Trans. Robot., vol. 31, no. 5, pp. 1147–1163, 2015

work page 2015

[2] [2]

Real-time loop closure in 2d LIDAR SLAM,

W. Hess, D. Kohler, H. Rapp, and D. Andor, “Real-time loop closure in 2d LIDAR SLAM,” in Proc. 2016 IEEE Int. Conf. Robot. Autom. (ICRA), Stockholm, Sweden, May 2016, pp. 1271–1278

work page 2016

[3] [3]

Illumination change robustness in direct visual SLAM,

S. Park, T. Schöps, and M. Pollefeys, “Illumination change robustness in direct visual SLAM,” in Proc. 2017 IEEE Int. Conf. Robot. Autom. (ICRA), Marina Bay Sands, Singapore, May 2017, pp. 4523–4530

work page 2017

[4] [4]

A compar- ative analysis of LiDAR SLAM-based indoor navigation for autonomous vehicles,

Q. Zou, Q. Sun, L. Chen, B. Nie, and Q. Li, “A compar- ative analysis of LiDAR SLAM-based indoor navigation for autonomous vehicles,” IEEE Trans. Intell. Transp. Syst., vol. 23, no. 7, pp. 6907–6921, 2022

work page 2022

[5] [5]

How feasible is the use of magnetic field alone for indoor positioning?

B. Li, T. Gallagher, A. G. Dempster, and C. Rizos, “How feasible is the use of magnetic field alone for indoor positioning?” in Proc. 2012 Int. Conf. Indoor Position. Indoor Navig. (IPIN), Sydney, Australia, Nov 2012, pp. 1–9

work page 2012

[6] [6]

Scalable magnetic field SLAM in 3D using gaussian process maps,

M. Kok and A. Solin, “Scalable magnetic field SLAM in 3D using gaussian process maps,” in Proc. 2018 21st Int. Conf. Inf. Fusion (FUSION), Cambridge, United Kingdom, July 2018, pp. 1353–1360

work page 2018

[7] [7]

An extended Kalman filter for magnetic field SLAM using Gaussian process regression,

F. Viset, R. Helmons, and M. Kok, “An extended Kalman filter for magnetic field SLAM using Gaussian process regression,” Sensors, vol. 22, no. 8, 2022

work page 2022

[8] [8]

Magnetic field-based SLAM method for solving the localization problem in mobile robot floor-cleaning task,

I. Vallivaara, J. Haverinen, A. Kemppainen, and J. Röning, “Magnetic field-based SLAM method for solving the localization problem in mobile robot floor-cleaning task,” in Proc. 2011 15th Int. Conf. Adv. Robot. (ICAR), Montevideo, Uruguay, June 2011, pp. 198–203

work page 2011

[9] [9]

Magnetic navigation using attitude-invariant magnetic field information for loop closure detection,

N. Pavlasek, C. C. Cossette, D. Roy-Guay, and J. R. Forbes, “Magnetic navigation using attitude-invariant magnetic field information for loop closure detection,” in Proc. 2023 IEEE/RSJ Int. Conf. Intell. Robots Syst. (IROS), Detroit, USA, 2023, pp. 5251–5257

work page 2023

[10] [10]

Saying goodbyes to rotating your phone: Magnetometer calibration during SLAM,

I. Vallivaara, Y. Dong, and T. Arslan, “Saying goodbyes to rotating your phone: Magnetometer calibration during SLAM,” in Proc. 2024 14th Int. Conf. Indoor Position. Indoor Navig. (IPIN), Hong Kong, China, Oct 2024, pp. 1–7

work page 2024

[11] [11]

MAINS: A magnetic-field-aided inertial navigation system for indoor positioning,

C. Huang, G. Hendeby, H. Fourati, C. Prieur, and I. Skog, “MAINS: A magnetic-field-aided inertial navigation system for indoor positioning,” IEEE Sens. J., vol. 24, no. 9, pp. 15 156– 15 166, 2024

work page 2024

[12] [12]

Signal architecture for a distributed magnetic local positioning system,

E. Prigge and J. How, “Signal architecture for a distributed magnetic local positioning system,” IEEE Sens. J., vol. 4, no. 6, pp. 864–873, 2004

work page 2004

[13] [13]

Orientation-aware 3d SLAM in alternating magnetic field from powerlines,

R. Wang, R. Tan, Z. Yan, and C. X. Lu, “Orientation-aware 3d SLAM in alternating magnetic field from powerlines,” Proc. ACM Interact. Mob. Wearable Ubiquitous Technol., vol. 7, no. 4, Jan. 2024

work page 2024

[14] [14]

IDF-MFL: Infrastructure-free and drift-free magnetic field localization for mobile robot,

H. Shen, Z. Wu, W. Wang, Q. Lyu, H. Zhou, and D. Wang, “IDF-MFL: Infrastructure-free and drift-free magnetic field localization for mobile robot,” in Proc. 2024 IEEE/RSJ Int. Conf. Intell. Robots Syst. (IROS), Abu Dhabi, UAE, October 2024, pp. 2278–2285

work page 2024

[15] [15]

Mag- odo: Motion speed estimation for indoor robots based on dual magnetometers,

T. Zhang, L. Wei, J. Kuang, H. Tang, and X. Niu, “Mag- odo: Motion speed estimation for indoor robots based on dual magnetometers,” Measurement, vol. 222, p. 113688, 2023

work page 2023

[16] [16]

Crowdmagmap: Crowdsourcing-based magnetic map construction for shopping mall,

Y. Wang, J. Kuang, T. Liu, X. Niu, and J. Liu, “Crowdmagmap: Crowdsourcing-based magnetic map construction for shopping mall,” IEEE Internet Things J., vol. 11, no. 3, pp. 5362–5373, 2024

work page 2024

[17] [17]

Crowdmagmap 2.0: Crowdsourced mag- netic mapping for multi-floor underground parking lot naviga- tion,

J. Kuang, Y. Wang, L. Ding, B. Zhou, L. Xu, L. Cao, L. He, Y. Wen, and X. Niu, “Crowdmagmap 2.0: Crowdsourced mag- netic mapping for multi-floor underground parking lot naviga- tion,” IEEE Trans. Intell. Transp. Syst., pp. 1–14, 2025

work page 2025

[18] [18]

Online one-dimensional magnetic field SLAM with loop-closure detection,

M. Kok and A. Solin, “Online one-dimensional magnetic field SLAM with loop-closure detection,” in Proc. 2024 IEEE Int. Conf. Multisensor Fusion Integr. Intell. Syst. (MFI), Pilsen, Czechia, Sep. 2024, pp. 1–7

work page 2024

[19] [19]

A pedestrian positioning method for urban canyon environments using magnetic field matching/inertial odometry fusion,

S. Chen, J. Kuang, Y. Wang, T. Wang, and X. Niu, “A pedestrian positioning method for urban canyon environments using magnetic field matching/inertial odometry fusion,” IEEE Trans. Instrum. Meas., vol. 74, pp. 1–15, 2025

work page 2025

[20] [20]

Magnetic odometry - a model-based approach using a sensor array,

I. Skog, G. Hendeby, and F. Gustafsson, “Magnetic odometry - a model-based approach using a sensor array,” in Proc. Int. Conf. Inf. Fusion (FUSION), Cambridge, United Kingdom, July 2018, pp. 794–798. 12

work page 2018

[21] [21]

Magnetic-field based odometry – an optical flow inspired approach,

I. Skog, G. Hendeby, and F. Trulsson, “Magnetic-field based odometry – an optical flow inspired approach,” in Proc. Int. Conf. Indoor Position. Indoor Navig. (IPIN), Lloret de Mar, Spain, Nov. 2021, pp. 1–8

work page 2021

[22] [22]

Magnetic Field Gradient- Based EKF for Velocity Estimation in Indoor Navigation,

M. Zmitri, H. Fourati, and C. Prieur, “Magnetic Field Gradient- Based EKF for Velocity Estimation in Indoor Navigation,” Sensors, vol. 20, no. 20, p. 5726, 2020

work page 2020

[23] [23]

Modeling and interpolation of the ambient magnetic field by Gaussian processes,

A. Solin, M. Kok, N. Wahlström, T. B. Schön, and S. Särkkä, “Modeling and interpolation of the ambient magnetic field by Gaussian processes,” IEEE Trans. Robot., vol. 34, no. 4, pp. 1112–1127, 2018

work page 2018

[24] [24]

Magnetic field norm SLAM using Gaussian process regression in foot-mounted sen- sors,

F. Viset, J. T. Gravdahl, and M. Kok, “Magnetic field norm SLAM using Gaussian process regression in foot-mounted sen- sors,” in Proc. Eur. Control Conf. (ECC), Rotterdam, Nether- lands, June 2021, pp. 392–398

work page 2021

[25] [25]

J. D. Jackson, Classical electrodynamics. John Wiley & Sons, 2021, vol. 2

work page 2021

[26] [26]

C. K. Williams and C. E. Rasmussen, Gaussian processes for machine learning. MIT press Cambridge, MA, 2006, vol. 2

work page 2006

[27] [28]

Quaternion kinematics for the error-state Kalman filter

[Online]. A vailable: http://arxiv.org/abs/1711.02508

work page internal anchor Pith review Pith/arXiv arXiv

[28] [29]

An observability- constrained magnetic field-aided inertial navigation system,

C. Huang, G. Hendeby, and I. Skog, “An observability- constrained magnetic field-aided inertial navigation system,” in Proc. 2024 14th Int. Conf. Indoor Position. Indoor Navig. (IPIN), Hong Kong, China, Oct. 2024, pp. 1–6. Chuan Huang (Student member, IEEE) re- ceived the B.Sc. from Beihang University in 2018 and the M.Sc. degree from China Electronics Te...

work page 2024