pith. machine review for the scientific record. sign in

arxiv: 2605.02054 · v1 · submitted 2026-05-03 · 📡 eess.SY · cs.CV· cs.RO· cs.SY

Recognition: 3 theorem links

· Lean Theorem

Observability Conditions and Filter Design for Visual Pose Estimation via Dual Quaternions

Nicholas B. Andrews , Kristi A. Morgansen
Authors on Pith no claims yet

Pith reviewed 2026-05-08 19:13 UTC · model grok-4.3

classification 📡 eess.SY cs.CVcs.ROcs.SY
keywords dual quaternionsvisual pose estimationunscented Kalman filterLie groupobservability analysis6-DOFPnP solverocclusions
0
0 comments X

The pith

A dual quaternion Lie group unscented Kalman filter models relative 6-DOF dynamics for improved visual pose estimation.

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

The paper develops a dual quaternion framework for 6-DOF visual target tracking to mitigate the noise sensitivity and inability to handle dropouts of traditional PnP solvers. It performs a nonlinear observability analysis using Lie algebraic methods to derive sufficient conditions for local observability with relative position vector and unit vector measurements. For unit vectors, the analysis recovers the collinear feature degeneracy as a rank deficiency. The authors then introduce a Lie group unscented Kalman filter that directly models relative dynamics without assuming cooperative measurements or slow motion. Simulations indicate this approach yields better pose accuracy and occlusion robustness than standard PnP methods, with applications to navigation and mapping.

Core claim

This paper claims that dual quaternions provide a natural representation for 6-DOF pose that, when combined with Lie group operations, permits a rigorous observability analysis and the design of an unscented Kalman filter capable of propagating estimates through measurement gaps while directly incorporating relative dynamics.

What carries the argument

The dual quaternion Lie group unscented Kalman filter, which uses the Lie group structure of dual quaternions to model and estimate 6-DOF relative pose and motion.

If this is right

  • The derived observability conditions ensure local observability for the specified sensing modalities.
  • The unit vector measurement case reproduces the classical collinear degeneracy via observability codistribution rank analysis.
  • The filter operates without assumptions on cooperative targets or slowly varying motion.
  • Simulations demonstrate superior accuracy and robustness to occlusions compared to PnP solvers.

Where Pith is reading between the lines

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

  • The method may integrate with inertial measurements to enhance visual-inertial navigation systems.
  • Observability insights could inform improvements in geometric PnP algorithms.
  • Direct dynamics modeling suggests potential for reduced drift in SLAM applications during visual interruptions.

Load-bearing premise

The dual quaternion representation and associated Lie group operations faithfully capture the 6-DOF relative dynamics without notable modeling errors or discretization problems.

What would settle it

A test case with high-speed relative motion and simulated occlusions where the proposed filter's error does not remain lower than the PnP solver's error would falsify the performance claim.

Figures

Figures reproduced from arXiv: 2605.02054 by Kristi A. Morgansen, Nicholas B. Andrews.

Figure 1
Figure 1. Figure 1: Diagram of the target (T), camera (C), and inertial (J) coordinate frames and the relative position vectors (rT /J , rC/J , rT /C ∈ R3 ) between them. Dual quaternion multiplication can be reconstructed in matrix form as aˆ ˆb = [ˆa]L ˆb = h ˆb i R a, ˆ (24a) aˆ ˆbcˆ = h aˆ ˆb i L cˆ = h ˆbcˆ i R aˆ = hˆb i R aˆ  cˆ = ˆa  [ˆc]R ˆb  (24b) for dual quaternions a, ˆ ˆb, cˆ ∈ DH. The dual quaternion cross … view at source ↗
Figure 2
Figure 2. Figure 2: 3-D plot showing the camera and target coordinate frame trajectories view at source ↗
Figure 3
Figure 3. Figure 3: Time history of the truth pose and velocities of the target relative view at source ↗
Figure 5
Figure 5. Figure 5: Pose estimation error comparison between DQ-UKF and OpenCV view at source ↗
read the original abstract

This paper presents a dual quaternion framework for 6-DOF visual target tracking that addresses key limitations of perspective-n-point (P$n$P) solvers: sensitivity to noise and outliers, and inability to propagate estimates through measurement dropouts. A nonlinear observability analysis is performed using a Lie algebraic approach, deriving sufficient conditions for local observability under two sensing modalities: relative position vector and unit vector measurements. For the unit vector case, the classical collinear feature point degeneracy of the perspective-three-point problem is recovered through rank analysis of the observability codistribution matrix, providing a control-theoretic interpretation of a previously geometric result. A dual quaternion Lie group unscented Kalman filter is then developed, directly modeling relative dynamics without assumptions about cooperative measurements or slowly-varying motion. Simulations demonstrate improved pose estimation accuracy and robustness to occlusions compared to an off-the-shelf P$n$P solver. Results are broadly applicable to visual-inertial navigation, simultaneous localization and mapping, and P$n$P solver development.

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 / 4 minor

Summary. The paper develops a dual-quaternion Lie-group framework for 6-DOF visual target tracking. It performs a nonlinear observability analysis via Lie-algebraic methods on the relative pose kinematics, deriving rank conditions for local observability under relative-position-vector and unit-vector measurements; the unit-vector case recovers the classical collinear degeneracy of the P3P problem. A dual-quaternion unscented Kalman filter on the Lie group is then constructed whose prediction step uses the dual-quaternion exponential map and does not invoke cooperative-target or slow-motion assumptions. Monte-Carlo simulations are reported to show improved pose accuracy and robustness to intermittent occlusions relative to an off-the-shelf PnP solver.

Significance. If the observability rank conditions and filter convergence claims hold under the stated sensing models, the work supplies a control-theoretic explanation of a well-known geometric degeneracy and a practical manifold filter that propagates estimates through measurement dropouts. These features are directly relevant to visual-inertial navigation, SLAM, and robust PnP pipelines. The explicit sigma-point construction on the dual-quaternion manifold and the parameter-free derivation of the observability codistribution are strengths that enhance reproducibility.

minor comments (4)
  1. [§3.2] §3.2, Eq. (17): the definition of the observability codistribution matrix is given, but the numerical rank computation procedure (tolerance, floating-point handling) is not stated; this affects verification of the reported rank drop for the collinear case.
  2. [§5] §5, Algorithm 1: the sigma-point generation on the dual-quaternion manifold is described, yet the retraction and inverse-retraction maps used for covariance propagation are only referenced; explicit formulas or a citation to the precise implementation would improve clarity.
  3. [§6] §6 Simulations: quantitative metrics (RMSE, NEES, or 3σ bounds) are mentioned but not tabulated for the occlusion scenarios; a table comparing the proposed filter against the PnP baseline across noise levels and dropout rates would strengthen the performance claim.
  4. Notation: the symbols for the dual-quaternion exponential map and the Lie-algebra projection operator are introduced without a consolidated table; readers unfamiliar with dual-quaternion conventions would benefit from an explicit symbol list.

Simulated Author's Rebuttal

0 responses · 0 unresolved

We thank the referee for the positive summary, significance assessment, and recommendation of minor revision for our manuscript on observability conditions and Lie-group filter design for dual-quaternion visual pose estimation. The recognition of the Lie-algebraic observability analysis and the manifold UKF's robustness to occlusions is appreciated. No specific major comments appear in the provided report, so we offer no point-by-point rebuttals below and will address any minor issues in the revision.

Circularity Check

0 steps flagged

No significant circularity; derivation is self-contained

full rationale

The paper's core chain consists of a Lie-algebraic observability analysis that produces explicit rank conditions on the codistribution matrix for two measurement modalities, recovering the known collinear degeneracy as a byproduct rather than assuming it. The dual-quaternion UKF is then constructed directly from the continuous-time kinematics and the exponential map on the Lie group, with sigma-point propagation stated explicitly. Neither step reduces to a fitted parameter renamed as a prediction, nor to a self-citation that bears the load of the central claim. Simulations supply independent empirical evidence of accuracy and occlusion robustness against a PnP baseline. All load-bearing steps are internally verifiable from the stated equations and standard manifold UKF construction once the kinematics are accepted; no self-definitional loop or ansatz smuggling is present.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 0 invented entities

Since only abstract available, limited details on any free parameters or axioms; the framework relies on standard Lie group theory and dual quaternion algebra which are established.

axioms (2)
  • domain assumption The system dynamics can be modeled using dual quaternion Lie group operations
    Assumed for the filter design
  • domain assumption Local observability conditions derived from Lie algebraic approach are sufficient for the application
    From the nonlinear observability analysis

pith-pipeline@v0.9.0 · 5478 in / 1248 out tokens · 24503 ms · 2026-05-08T19:13:34.567206+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

43 extracted references · 3 canonical work pages

  1. [1]

    EPnP: An accurate O(n) solution to the PnP problem,

    V . Lepetit, F. Moreno-Noguer, and P. Fua, “EPnP: An accurate O(n) solution to the PnP problem,”Int. J. Comput. Vis., vol. 81, no. 2, pp. 155–166, Feb. 2009

    2009

  2. [2]

    An efficient algebraic solution to the perspective-three-point problem,

    T. Ke and S. I. Roumeliotis, “An efficient algebraic solution to the perspective-three-point problem,” in2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Jul. 2017

    2017

  3. [3]

    A consistently fast and globally optimal solution to the perspective-n-point problem,

    G. Terzakis and M. Lourakis, “A consistently fast and globally optimal solution to the perspective-n-point problem,” inLecture Notes in Computer Science, ser. Lecture Notes in Computer Science. Cham: Springer International Publishing, 2020, pp. 478–494

    2020

  4. [4]

    Dual quaternion particle filtering for pose estimation,

    A. Sveier and O. Egeland, “Dual quaternion particle filtering for pose estimation,”IEEE Trans. Control Syst. Technol., vol. 29, no. 5, pp. 2012–2025, Sep. 2021

    2012

  5. [5]

    EKFPnP: extended kalman filter for camera pose estimation in a sequence of images,

    M. A. Mehralian and M. Soryani, “EKFPnP: extended kalman filter for camera pose estimation in a sequence of images,”IET Image Process., vol. 14, no. 15, pp. 3774–3780, Dec. 2020

    2020

  6. [6]

    A computational analysis of screw trans- formations in robotics,

    J. Funda and R. P. Paul, “A computational analysis of screw trans- formations in robotics,”IEEE Trans. Rob. Autom., vol. 6, no. 3, pp. 348–356, Jun. 1990

    1990

  7. [7]

    Relative pose observability analysis using dual quaternions,

    N. B. Andrews and K. A. Morgansen, “Relative pose observability analysis using dual quaternions,” in2024 IEEE 63rd Conference on Decision and Control (CDC), Dec. 2024, pp. 7740–7745

    2024

  8. [8]

    Coupled 6-dof control for distributed aerospace systems,

    T. P. Reynolds and M. Mesbahi, “Coupled 6-dof control for distributed aerospace systems,” in2018 IEEE Conference on Decision and Control (CDC), 2018, pp. 5294–5299

    2018

  9. [9]

    Adaptive position and attitude-tracking controller for satellite proximity operations using dual quaternions,

    N. Filipe and P. Tsiotras, “Adaptive position and attitude-tracking controller for satellite proximity operations using dual quaternions,” J. Guid. Control Dyn., vol. 38, no. 4, pp. 566–577, Apr. 2015

    2015

  10. [10]

    Dis- tributed dual-quaternion extended kalman filtering for spacecraft pose estimation,

    M. Hudoba de Badyn, J. Binz, A. Iannelli, and R. S. Smith, “Dis- tributed dual-quaternion extended kalman filtering for spacecraft pose estimation,”J. Guid. Control Dyn., vol. 48, no. 5, pp. 1071–1087, May 2025

    2025

  11. [11]

    Relative navigation in asteroid missions using dual quaternion filtering,

    B. Razgus, E. Mooij, and D. Choukroun, “Relative navigation in asteroid missions using dual quaternion filtering,”J. Guid. Control Dyn., vol. 40, no. 9, pp. 2151–2166, Sep. 2017

    2017

  12. [12]

    Extended kalman filter for spacecraft pose estimation using dual quaternions,

    N. Filipe, M. Kontitsis, and P. Tsiotras, “Extended kalman filter for spacecraft pose estimation using dual quaternions,”J. Guid. Control Dyn., vol. 38, no. 9, pp. 1625–1641, Sep. 2015

    2015

  13. [13]

    Dual quaternion kalman filtering and observability analysis for satellite relative navigation with line-of-sight measurements,

    Y . Zivan and D. Choukroun, “Dual quaternion kalman filtering and observability analysis for satellite relative navigation with line-of-sight measurements,”IEEE Trans. Aerosp. Electron. Syst., vol. 58, no. 2, pp. 754–765, Apr. 2022

    2022

  14. [14]

    Riccati observers for the nonstationary PnP problem,

    T. Hamel and C. Samson, “Riccati observers for the nonstationary PnP problem,”IEEE Trans. Automat. Contr., vol. 63, no. 3, pp. 726–741, Mar. 2018

    2018

  15. [15]

    Relative pose estimation from bearing measurements of three unknown source points,

    M.-D. Hua, S. De Marco, T. Hamel, and R. W. Beard, “Relative pose estimation from bearing measurements of three unknown source points,” in2020 59th IEEE Conference on Decision and Control (CDC), Dec. 2020, pp. 4176–4181

    2020

  16. [16]

    Observability analysis of six-degree- of-freedom configuration determination using vector observations,

    D. Sun and J. L. Crassidis, “Observability analysis of six-degree- of-freedom configuration determination using vector observations,”J. Guid. Control Dyn., vol. 25, no. 6, pp. 1149–1157, Nov. 2002

    2002

  17. [17]

    A kalman filter-based algorithm for IMU-camera calibration: Observability analysis and performance evaluation,

    F. M. Mirzaei and S. I. Roumeliotis, “A kalman filter-based algorithm for IMU-camera calibration: Observability analysis and performance evaluation,”IEEE Trans. Robot., vol. 24, no. 5, pp. 1143–1156, Oct. 2008

    2008

  18. [18]

    Unscented kalman filtering on lie groups,

    M. Brossard, S. Bonnabel, and J.-P. Condomines, “Unscented kalman filtering on lie groups,” in2017 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), Sep. 2017

    2017

  19. [19]

    A practical exploration of a dual quaternion based unscented kalman filter for pose estimation,

    T. N. Zambo, “A practical exploration of a dual quaternion based unscented kalman filter for pose estimation,” Ph.D. dissertation, Delft University of Technology, Dec. 2024

    2024

  20. [20]

    Spacecraft robot kinematics using dual quaternions,

    A. Valverde and P. Tsiotras, “Spacecraft robot kinematics using dual quaternions,”Robotics, vol. 7, no. 4, p. 64, Oct. 2018

    2018

  21. [21]

    Quaternion kinematics for the error-state kalman filter,

    J. Sol `a, “Quaternion kinematics for the error-state kalman filter,”

  22. [22]

    Quaternion kinematics for the error-state Kalman filter

    [Online]. Available: https://arxiv.org/abs/1711.02508

    work page Pith review arXiv

  23. [23]

    Dual-quaternion analytic lqr control design for spacecraft proximity operations,

    K. Stanfield and A. Bani Younes, “Dual-quaternion analytic lqr control design for spacecraft proximity operations,”Sensors, vol. 21, no. 11, 2021

    2021

  24. [24]

    6-DOF robust adaptive terminal sliding mode control for spacecraft formation flying,

    J. Wang and Z. Sun, “6-DOF robust adaptive terminal sliding mode control for spacecraft formation flying,”Acta Astronaut., vol. 73, no. C, pp. 76–87, Apr. 2012

    2012

  25. [25]

    Rigid body motion tracking without linear and angular velocity feedback using dual quaternions,

    N. Filipe and P. Tsiotras, “Rigid body motion tracking without linear and angular velocity feedback using dual quaternions,” in2013 European Control Conference (ECC), Jul. 2013, pp. 329–334

    2013

  26. [26]

    Nonlinear controllability and observabil- ity,

    R. Hermann and A. Krener, “Nonlinear controllability and observabil- ity,”IEEE Trans. Automat. Contr., vol. 22, no. 5, pp. 728–740, Oct. 1977

    1977

  27. [27]

    Nijmeijer and A

    H. Nijmeijer and A. J. van der Schaft,Nonlinear Dynamical Control Systems. Springer Science & Business Media, Apr. 1990

    1990

  28. [28]

    Rank of a lower triangular block matrix,

    N. Alger, “Rank of a lower triangular block matrix,” Mathematics Stack Exchange. [Online]. Available: https://math.stackexchange.com/ q/1867025

    work page arXiv

  29. [29]

    Optimal fiducial marker placement for satellite proximity operations using observability gramians,

    N. B. Andrews and K. A. Morgansen, “Optimal fiducial marker placement for satellite proximity operations using observability gramians,” 2025. [Online]. Available: https://arxiv.org/abs/2501.01704

    work page arXiv 2025

  30. [30]

    Unscented dual quaternion particle filter for SE(3) estimation,

    K. Li, F. Pfaff, and U. D. Hanebeck, “Unscented dual quaternion particle filter for SE(3) estimation,”IEEE Control Syst. Lett., vol. 5, no. 2, pp. 647–652, Apr. 2021

    2021

  31. [31]

    Schmidt-ekf- based visual-inertial moving object tracking,

    K. Eckenhoff, P. Geneva, N. Merrill, and G. Huang, “Schmidt-ekf- based visual-inertial moving object tracking,” in2020 IEEE Inter- national Conference on Robotics and Automation (ICRA), 2020, pp. 651–657

    2020

  32. [32]

    Tightly-coupled visual-inertial localization and 3-D rigid-body target tracking,

    K. Eckenhoff, Y . Yang, P. Geneva, and G. Huang, “Tightly-coupled visual-inertial localization and 3-D rigid-body target tracking,”IEEE Robot. Autom. Lett., vol. 4, no. 2, pp. 1541–1548, Apr. 2019

    2019

  33. [33]

    Geometric integration of quater- nions,

    M. S. Andrle and J. L. Crassidis, “Geometric integration of quater- nions,”Journal of Guidance, Control, and Dynamics, vol. 36, no. 6, pp. 1762–1767, 2013

    2013

  34. [34]

    Attitude determination from vector observations: Quaternion estimation,

    I. Y . Bar-Itzhack and Y . Oshman, “Attitude determination from vector observations: Quaternion estimation,”IEEE Transactions on Aerospace and Electronic Systems, no. 1, pp. 128–136, 2007

    2007

  35. [35]

    Quaternion normalization in additive EKF for spacecraft attitude determination,

    I. Y . Bar-Itzhack, J. Deutschmann, and F. L. Markley, “Quaternion normalization in additive EKF for spacecraft attitude determination,” inNavigation and Control Conference, Aug. 1991

    1991

  36. [36]

    Open source computer vision library,

    OpenCV , “Open source computer vision library,” https://opencv.org/

  37. [37]

    Revisiting the pnp problem: A fast, general and optimal solution,

    Y . Zheng, Y . Kuang, S. Sugimoto, K. Astrom, and M. Okutomi, “Revisiting the pnp problem: A fast, general and optimal solution,” inProceedings of the IEEE International Conference on Computer Vision (ICCV), Dec. 2013

    2013

  38. [38]

    Very fast solution to the PnP problem with algebraic outlier rejection,

    L. Ferraz, X. Binefa, and F. Moreno-Noguer, “Very fast solution to the PnP problem with algebraic outlier rejection,” in2014 IEEE Conference on Computer Vision and Pattern Recognition, Jun. 2014

    2014

  39. [39]

    Relative computer vision-based navigation for small inspection spacecraft,

    B. E. Tweddle and A. Saenz-Otero, “Relative computer vision-based navigation for small inspection spacecraft,”J. Guid. Control Dyn., vol. 38, no. 5, pp. 969–978, May 2015

    2015

  40. [40]

    Gyroscopic sensing in the wings of the hawkmoth manduca sexta: the role of sensor location and directional sensitivity,

    B. T. Hinson and K. A. Morgansen, “Gyroscopic sensing in the wings of the hawkmoth manduca sexta: the role of sensor location and directional sensitivity,”Bioinspir. Biomim., vol. 10, no. 5, p. 056013, Oct. 2015

    2015

  41. [41]

    Sensor placement on a cantilever beam using observability gramians,

    N. L. Brace, N. B. Andrews, J. Upsal, and K. A. Morgansen, “Sensor placement on a cantilever beam using observability gramians,” in2022 IEEE 61st Conference on Decision and Control (CDC), 2022, pp. 388– 395

    2022

  42. [42]

    Observability-informed optimal sensor placement for soft robots,

    S. Smocot, “Observability-informed optimal sensor placement for soft robots,” Master’s thesis, McGill University, Montreal, 2026

    2026

  43. [43]

    The unscented kalman filter for non- linear estimation,

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

    2000