pith. sign in

arxiv: 2509.04016 · v1 · submitted 2025-09-04 · 💻 cs.RO

Odometry Calibration and Pose Estimation of a 4WIS4WID Mobile Wall Climbing Robot

Branimir \'Caran , Vladimir Mili\'c , Marko \v{S}vaco , Bojan Jerbi\'c This is my paper

Pith reviewed 2026-05-18 19:23 UTC · model grok-4.3

classification 💻 cs.RO
keywords wall climbing robotodometry calibrationpose estimation4WIS4WIDvisual odometryIMUKalman filter
0
0 comments X

The pith

Wall-climbing robots track position on facades by calibrating wheel parameters and fusing wheel odometry, visual odometry, and IMU data with EKF and UKF.

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

This paper develops a pose estimator for a four-wheel independently steered and driven wall-climbing robot that operates on building surfaces. It first tunes the robot's kinematic parameters with nonlinear optimization and evolutionary search methods, then combines wheel speeds, camera-derived motion, and inertial readings inside two different Kalman filters. The work matters because these robots must carry measurement tools and perform maintenance where GPS signals fade and line-of-sight sensors cannot be used reliably. Accurate pose data lets the robot know where it is at each moment of inspection or repair. Experiments on a physical prototype compare the calibration routines and filter outputs to show how drift is managed in practice.

Core claim

The paper establishes that systematic calibration of the 4WIS4WID kinematic model using Levenberg-Marquardt, genetic algorithms, and particle swarm optimization, followed by real-time fusion of wheel odometry, visual odometry, and IMU measurements in both EKF and UKF, produces usable pose estimates for vertical-surface operation.

What carries the argument

Multimodal sensor fusion through Extended Kalman Filter and Unscented Kalman Filter applied after kinematic parameter calibration of the 4WIS4WID drive system.

If this is right

  • Calibrated wheel parameters reduce systematic velocity errors during wall motion.
  • EKF and UKF outputs can be compared directly in the same experimental runs to assess filter behavior.
  • The system supports carrying precise measurement tools by supplying continuous pose information during facade work.
  • Both deterministic and stochastic calibration methods yield usable parameter sets for the robot.

Where Pith is reading between the lines

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

  • The same calibration-plus-fusion pattern could be tested on other mobile platforms that lose GPS in cluttered or indoor settings.
  • Periodic re-calibration may be needed if wheel diameters change through wear during extended deployments.
  • If a facade has large uniform areas with few visual features, the filter would have to increase reliance on the inertial channel to stay accurate.

Load-bearing premise

The assumption that visual features on the facade stay sufficiently distinct and that wheel contact remains consistent enough for the fused measurements to keep drift bounded.

What would settle it

Record the robot's estimated pose while it follows a measured path on a vertical test wall and compare it against an independent external tracking system; large cumulative position error at the end of the path would show the estimator fails to control drift.

Figures

Figures reproduced from arXiv: 2509.04016 by Bojan Jerbi\'c, Branimir \'Caran, Marko \v{S}vaco, Vladimir Mili\'c.

Figure 1
Figure 1. Figure 1: Computer aided design of the wall robot. [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: 4WIS4WID wall climbing robot driving on vertical surface. [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: 4WIS4WID mobile robot kinematic structure. [PITH_FULL_IMAGE:figures/full_fig_p003_3.png] view at source ↗
Figure 5
Figure 5. Figure 5: Circular movement in CCW (top) and CW (bottom) directions. [PITH_FULL_IMAGE:figures/full_fig_p004_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Rotational movement around robot’s z axis in CCW(top) and CW(bottom) direction. z that minimizes the sum of squares function subject to possible constraints. In this case, the calibration problem has one inequality constraints of upper and lower bounds for the sought vector. It can be concluded that the calibration problem described belongs to the subset of problems that these methods address. C. Odometry … view at source ↗
Figure 7
Figure 7. Figure 7: Calibration methods comparison for straight line along [PITH_FULL_IMAGE:figures/full_fig_p005_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: Calibration methods comparison for circular movement in clockwise [PITH_FULL_IMAGE:figures/full_fig_p006_8.png] view at source ↗
Figure 10
Figure 10. Figure 10: Wall climbing robot control systems design. [PITH_FULL_IMAGE:figures/full_fig_p006_10.png] view at source ↗
Figure 11
Figure 11. Figure 11: Experimental setup for pose estimator comparison. [PITH_FULL_IMAGE:figures/full_fig_p007_11.png] view at source ↗
Figure 12
Figure 12. Figure 12: Experimental results for pose estimators in [PITH_FULL_IMAGE:figures/full_fig_p007_12.png] view at source ↗
Figure 13
Figure 13. Figure 13: Experimental results for pose estimators in [PITH_FULL_IMAGE:figures/full_fig_p008_13.png] view at source ↗
read the original abstract

This paper presents the design of a pose estimator for a four wheel independent steer four wheel independent drive (4WIS4WID) wall climbing mobile robot, based on the fusion of multimodal measurements, including wheel odometry, visual odometry, and an inertial measurement unit (IMU) data using Extended Kalman Filter (EKF) and Unscented Kalman Filter (UKF). The pose estimator is a critical component of wall climbing mobile robots, as their operational environment involves carrying precise measurement equipment and maintenance tools in construction, requiring information about pose on the building at the time of measurement. Due to the complex geometry and material properties of building facades, the use of traditional localization sensors such as laser, ultrasonic, or radar is often infeasible for wall-climbing robots. Moreover, GPS-based localization is generally unreliable in these environments because of signal degradation caused by reinforced concrete and electromagnetic interference. Consequently, robot odometry remains the primary source of velocity and position information, despite being susceptible to drift caused by both systematic and non-systematic errors. The calibrations of the robot's systematic parameters were conducted using nonlinear optimization and Levenberg-Marquardt methods as Newton-Gauss and gradient-based model fitting methods, while Genetic algorithm and Particle swarm were used as stochastic-based methods for kinematic parameter calibration. Performance and results of the calibration methods and pose estimators were validated in detail with experiments on the experimental mobile wall climbing robot.

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.

Referee Report

0 major / 2 minor

Summary. This paper presents the design of a pose estimator for a 4WIS4WID wall-climbing mobile robot that fuses wheel odometry, visual odometry, and IMU measurements using both EKF and UKF. It additionally describes calibration of the robot's systematic kinematic parameters via Levenberg-Marquardt optimization as well as stochastic methods (genetic algorithms and particle-swarm optimization), with experimental validation performed on the physical platform that reports quantitative RMSE reductions relative to wheel-odometry-only baselines.

Significance. If the reported trajectory errors and calibration results hold, the work supplies a concrete, experimentally grounded solution for pose estimation on building facades where GPS and conventional ranging sensors are impractical. The explicit statement of kinematic models, measurement equations, calibration objective functions, and the comparison of fused estimates against a wheel-odometry baseline constitute a reproducible contribution that can be directly tested on similar 4WIS4WID platforms.

minor comments (2)
  1. The abstract would be strengthened by including at least one quantitative result (e.g., RMSE reduction achieved by the fused estimator) so that the performance claim is immediately visible to readers.
  2. A direct side-by-side comparison of EKF versus UKF accuracy, computational cost, and robustness on the same trajectories would help readers decide which filter is preferable for this robot class.

Simulated Author's Rebuttal

0 responses · 0 unresolved

We thank the referee for the positive summary of our manuscript on multimodal pose estimation and kinematic calibration for the 4WIS4WID wall-climbing robot, as well as the recommendation for minor revision. The referee's description accurately captures the EKF/UKF fusion approach, calibration methods (Levenberg-Marquardt, genetic algorithms, and particle swarm optimization), and experimental validation against wheel-odometry baselines.

Circularity Check

0 steps flagged

No significant circularity; standard sensor fusion and calibration on experimental data

full rationale

The manuscript applies textbook EKF/UKF fusion to wheel odometry, visual odometry and IMU measurements for a 4WIS4WID wall-climbing robot, together with explicit kinematic calibration via Levenberg-Marquardt, genetic algorithms and particle-swarm optimization. All model equations, measurement models, objective functions and reported RMSE values are stated directly; the results are obtained by running the filters and optimizers on physical trajectory data. No derivation step equates a claimed prediction to a fitted parameter by construction, no load-bearing uniqueness theorem is imported via self-citation, and no ansatz is smuggled through prior work. The central claims therefore remain independent of the inputs they are evaluated against.

Axiom & Free-Parameter Ledger

1 free parameters · 1 axioms · 0 invented entities

The central claim depends on standard sensor noise models and the kinematic model of the 4WIS4WID configuration; calibration treats systematic parameters as adjustable quantities fitted to data.

free parameters (1)
  • systematic kinematic parameters
    Wheel and steering parameters calibrated via nonlinear optimization and stochastic methods to correct odometry errors.
axioms (1)
  • domain assumption Sensor noise in wheel odometry, visual odometry, and IMU can be adequately modeled for EKF and UKF application.
    Implicit in the choice of Kalman filters for multimodal fusion as stated in the abstract.

pith-pipeline@v0.9.0 · 5805 in / 1423 out tokens · 60601 ms · 2026-05-18T19:23:29.252203+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Lean theorems connected to this paper

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

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.

Reference graph

Works this paper leans on

29 extracted references · 29 canonical work pages

  1. [1]

    A survey of wall climbing robots: Recent advances and challenges,

    S. Nansai and R. E. Mohan, “A survey of wall climbing robots: Recent advances and challenges,” Robotics, vol. 5, p. 14, 2016

    work page 2016

  2. [2]

    Climbing robots for maintenance and inspections of vertical structures—a survey of design aspects and tech- nologies,

    D. Schmidt and K. Berns, “Climbing robots for maintenance and inspections of vertical structures—a survey of design aspects and tech- nologies,” Robotics and Autonomous Systems , vol. 61, pp. 1288–1305, 2013

    work page 2013

  3. [3]

    Advances in climbing robots for vertical structures in the past decade: A review,

    G. Fang and J. Cheng, “Advances in climbing robots for vertical structures in the past decade: A review,” Biomimetics, vol. 8, p. 47, 2023

    work page 2023

  4. [4]

    Design and technical development of wall-climbing robots: A review,

    Y . Fang, S. Wang, Q. Bi, D. Cui, and C. Yan, “Design and technical development of wall-climbing robots: A review,” Journal of Bionic Engineering, vol. 19, 2022

    work page 2022

  5. [5]

    Dynamics of a four wheeled wall climbing robot,

    A. Chokshi and J. Mehta, “Dynamics of a four wheeled wall climbing robot,” in Advances in Mechanical Engineering , 2021, pp. 530–536

    work page 2021

  6. [6]

    Kinematics, dynamics and control design of 4wis4wid mobile robots,

    T.-H. Li and M.-H. Lee, “Kinematics, dynamics and control design of 4wis4wid mobile robots,” The Journal of Engineering , vol. 2015, 2014

    work page 2015

  7. [7]

    Mobile robot localization: Current challenges and future prospective,

    I. Ullah, D. Adhikari, H. Khan, M. S. Anwar, S. Ahmad, and X. Bai, “Mobile robot localization: Current challenges and future prospective,” Computer Science Review , vol. 53, p. 100651, 2024

    work page 2024

  8. [8]

    Climbing robots for manufacturing,

    B. Tao, Z. Gong, and H. Ding, “Climbing robots for manufacturing,” National Science Review , vol. 10, p. nwad042, 2023

    work page 2023

  9. [9]

    Review of simultaneous localization and mapping (slam) for construction robotics applications,

    A. Yarovoi and Y . K. Cho, “Review of simultaneous localization and mapping (slam) for construction robotics applications,” Automation in Construction, vol. 162, p. 105344, 2024

    work page 2024

  10. [10]

    Information fusion of gps, ins and odometer sensors for improving localization accuracy of mobile robots in indoor and outdoor applications,

    S. Yousuf and M. B. Kadri, “Information fusion of gps, ins and odometer sensors for improving localization accuracy of mobile robots in indoor and outdoor applications,” Robotica, vol. 39, pp. 250–276, 2021

    work page 2021

  11. [11]

    Visual positioning of distant wall-climbing robots using convolutional neural networks,

    Q. Zhou and X. Li, “Visual positioning of distant wall-climbing robots using convolutional neural networks,” Journal of Intelligent & Robotic Systems, vol. 98, 2020

    work page 2020

  12. [12]

    An autonomous localisation method of wall-climbing robots on large steel components with imu and fixed rgb-d camera,

    W. Zhang, T. Huang, and Z. Sun, “An autonomous localisation method of wall-climbing robots on large steel components with imu and fixed rgb-d camera,” Journal of Navigation , pp. 1–16, 2025

    work page 2025

  13. [13]

    Position- tracking control of dual-rope winch robot with rope slip compensation,

    S. Yoo, T. Kim, M. Seo, J. Oh, H. S. Kim, and T. Seo, “Position- tracking control of dual-rope winch robot with rope slip compensation,” IEEE/ASME Transactions on Mechatronics , vol. 26, pp. 1754–1762, 2021

    work page 2021

  14. [14]

    Development of a mobile wall- climbing robot with a hybrid adhesion system,

    M. Bo ˇzi´c, B. Jerbi ´c, and M. ˇSvaco, “Development of a mobile wall- climbing robot with a hybrid adhesion system,” in 2021 44th Inter- national Convention on Information and Communication Technology, Electronics and Microelectronics (MIPRO), 2021

    work page 2021

  15. [15]

    Mobile wall- climbing robot for ndt inspection of vertical concrete structures,

    M. Bo ˇzi´c, B. ´Caran, M. ˇSvaco, B. Jerbi ´c, and M. Serdar, “Mobile wall- climbing robot for ndt inspection of vertical concrete structures,” in e-Journal of Nondestructive Testing , vol. 27, no. 9, 2022

    work page 2022

  16. [16]

    Modeling, parameter estimation and control design of 4wis4wid mobile robot: Simulation and experimental validation,

    B. ´Caran, M. ˇSvaco, F. ˇSuligoj, and B. Jerbi ´c, “Modeling, parameter estimation and control design of 4wis4wid mobile robot: Simulation and experimental validation,” inAdvances in Service and Industrial Robotics, 2024, pp. 607–618

    work page 2024

  17. [17]

    Evolution of odometry cal- ibration methods for ground mobile robots,

    R. B. Sousa, M. R. Petry, and A. P. Moreira, “Evolution of odometry cal- ibration methods for ground mobile robots,” in 2020 IEEE International Conference on Autonomous Robot Systems and Competitions (ICARSC) , 2020, pp. 294–299

    work page 2020

  18. [18]

    Kalman filter: Historical overview and review of its use in robotics 60 years after its creation,

    C. Urrea and R. Agramonte, “Kalman filter: Historical overview and review of its use in robotics 60 years after its creation,” Journal of Sensors, pp. 1–21, 2021

    work page 2021

  19. [19]

    Optimal control-based algorithm design and application for trajectory tracking of a mobile robot with four independently steered and four independently actuated wheels,

    B. ´Caran, V . Mili´c, M. ˇSvaco, and B. Jerbi ´c, “Optimal control-based algorithm design and application for trajectory tracking of a mobile robot with four independently steered and four independently actuated wheels,” Actuators, vol. 13, p. 279, 2024

    work page 2024

  20. [20]

    Mobile robot localization via ekf and ukf: A comparison based on real data,

    L. D’Alfonso, W. Lucia, P. Muraca, and P. Pugliese, “Mobile robot localization via ekf and ukf: A comparison based on real data,” Robotics and Autonomous Systems , pp. 122–127, 2015

    work page 2015

  21. [21]

    Unscented filtering and nonlinear estimation,

    S. J. Julier and J. K. Uhlmann, “Unscented filtering and nonlinear estimation,” in Proceedings of the IEEE , 2004, p. 401 – 422

    work page 2004

  22. [22]

    The unscented kalman filter for non- linear estimation,

    E. Wan and R. Van Der Merwe, “The unscented kalman filter for non- linear estimation,” in Proceedings of the IEEE 2000 Adaptive Systems for Signal Processing, Communications, and Control Symposium (Cat. No.00EX373), 2000, pp. 153–158

    work page 2000

  23. [23]

    Unscented kalman filtering for additive noise case: Augmented versus nonaugmented,

    Y . Wu, D. Hu, M. Wu, and X. Hu, “Unscented kalman filtering for additive noise case: Augmented versus nonaugmented,” IEEE Signal Processing Letters, vol. 12, no. 5, p. 357 – 360, 2005. IEEE EUROPEAN CONFERENCE ON MOBILE ROBOTS 2025. PREPRINT VERSION. ACCEPTED JUNE, 2025

    work page 2005

  24. [24]

    Measurement and correction of systematic odometry errors in mobile robots,

    J. Borenstein and L. Feng, “Measurement and correction of systematic odometry errors in mobile robots,” IEEE Transactions on Robotics and Automation, pp. 869–880, 1996

    work page 1996

  25. [25]

    A calibration method for odometry of mobile robots based on the least-squares technique: theory and experimental validation,

    G. Antonelli, S. Chiaverini, and G. Fusco, “A calibration method for odometry of mobile robots based on the least-squares technique: theory and experimental validation,” IEEE Transactions on Robotics , pp. 994– 1004, 2005

    work page 2005

  26. [26]

    Levenberg–marquardt methods with strong local convergence properties for solving nonlinear equations with convex constraints,

    C. Kanzow, N. Yamashita, and M. Fukushima, “Levenberg–marquardt methods with strong local convergence properties for solving nonlinear equations with convex constraints,” Journal of Computational and Applied Mathematics, pp. 375–397, 2004

    work page 2004

  27. [27]

    An interior- point method for large-scale ℓ1-regularized least squares,

    S.-J. Kim, K. Koh, M. Lustig, S. Boyd, and D. Gorinevsky, “An interior- point method for large-scale ℓ1-regularized least squares,” IEEE Journal of Selected Topics in Signal Processing , pp. 606–617, 2007

    work page 2007

  28. [28]

    Non-parametric calibra- tion of the inverse kinematic matrix of a three-wheeled omnidirectional mobile robot based on genetic algorithms,

    J. Palac ´ın, E. Rubies, R. Bitri`a, and E. Clotet, “Non-parametric calibra- tion of the inverse kinematic matrix of a three-wheeled omnidirectional mobile robot based on genetic algorithms,” Applied Sciences , p. 1053, 2023

    work page 2023

  29. [29]

    Particle swarm optimization,

    J. Kennedy and R. Eberhart, “Particle swarm optimization,” in Proceed- ings of ICNN’95 - International Conference on Neural Networks , 1995, pp. 1942–1948

    work page 1995