Pith. sign in

REVIEW 23 references

Attitude Estimation Using Scalar Measurements

T0 review · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read Full 3D attitude can be estimated from scalar measurements alone.

desk verdict A useful LTV generalization of attitude estimation to scalar measurements with a correct observability lemma, but the GES proof covers the reset-free filter, not the implemented SVD-reset algorithm. read the letter →

arxiv 2506.07068 v1 pith:LQBMZVHT submitted 2025-06-08 eess.SY cs.SY

Hassan Alnahhal , Sifeddine Benahmed , Soulaimane Berkane , Tarel Hamel This is my paper
classification eess.SYcs.SY MSC 93B0793C0593D20
keywords attitudeestimationscalarmeasurementslineartime-varyingsystemsuniformobservabilitydeterministicKalmanfilterglobalexponentialstabilitypartialvectorpersistentexcitation
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

This paper claims that a rigid body's full three-dimensional orientation can be estimated from scalar measurements alone — one surviving axis of a vector sensor, a Pitot tube, a barometer merged with a range finder, a landmark height difference — with global exponential convergence rather than a local guarantee. The construction stacks the columns of the transposed rotation matrix into a nine-dimensional vector, so the nonlinear rotation kinematics become a linear time-varying system and every scalar constraint $y_i=a_i^\top R^\top b_i$ becomes a linear output. The authors prove the pair is uniformly observable exactly when the persistent-excitation integral in Lemma 1, involving $u_i=b_i\otimes R a_i$, is uniformly positive definite, and that under this condition the deterministic Kalman filter is globally uniformly exponentially stable. The practical payoff is that an estimator no longer needs full directional information to lock onto attitude, and it can keep working when some sensor axes fail.

What carries the argument

The central object is the stacked state $x_B=\operatorname{vec}(R^\top)$, which carries the argument by converting nonlinear rotation kinematics into the linear time-varying equation $\dot x_B=-(I_3\otimes[\omega]_\times)x_B$. Each scalar measurement is then the linear output $y_i=(b_i^\top\otimes a_i^\top)x_B$. Lemma 1 is the load-bearing identity: writing $T(s)=I_3\otimes R(s)$, the observability Gramian factorizes as $W=T^\top\bar W T$, so uniform observability is equivalent to the integral in (19) being uniformly positive definite. The deterministic Kalman gain, with the Riccati equation, supplies the Lyapunov function $V=\tilde x_B^\top P^{-1}\tilde x_B$, whose derivative is negative definite under (19), yielding the global exponential stability claim.

What would settle it

Simulate the observer on a trajectory whose angular velocity is always parallel to $e_3$ and feed it a single scalar measurement with both $a_i$ and $b_i$ parallel to $e_3$; then $u_i(s)=b_i\otimes R(s)a_i$ is constant, the integral in (19) has rank one, the pair is not uniformly observable, and the filter will fail to reduce estimation error in the orthogonal directions.

Watch

Extended reading notes

Core claim

On the paper's own terms, the central discovery is an exact reformulation: the attitude kinematics $\dot R=R[\omega]_\times$ with scalar measurements $y_i=a_i^\top R^\top b_i$ become the nine-dimensional linear time-varying system (13) with state $x_B=\operatorname{vec}(R^\top)$ and output rows $b_i^\top\otimes a_i^\top$. Lemma 1 then characterizes uniform observability of $(A(\cdot),C(\cdot))$ as the persistent-excitation condition (19), and the deterministic linear Kalman filter (14) with gain $K=PC^\top Q$ therefore enjoys global uniform exponential stability whenever (19) holds. No linearization is involved; the only nonlinear operation is the optional SVD projection (17) back onto $\mathrm{SO}(3)$, which the paper shows locally reduces or preserves the distance to the true rotation.

Load-bearing premise

The convergence guarantee rests on the persistent-excitation condition (19), which depends on the true rotation and must hold along the actual trajectory; the paper gives no explicit maneuver or sensor-geometry conditions for the Pitot, barometer/range, and landmark modalities to satisfy it, and it also assumes negligible wind, flat ground, and known landmark height differences.

Editorial extensions

If this is right

  • A vehicle whose magnetometer or accelerometer has a dead axis can still get a full-attitude estimate from the surviving scalar components, provided the remaining constraints persistently excite the state.
  • Sensors like a Pitot tube, a barometer with a down-facing range finder, and landmark vertical-offset constraints can be fused in the same filter as vector measurements, with no linearization.
  • The stability guarantee is global and exponential under (19), so the estimator recovers from arbitrarily large initial attitude errors whenever the condition holds.
  • The closed-form discrete implementation with zero-order hold and per-sensor update rates allows asynchronous IMU and scalar-sensor fusion in a straightforward way.
  • Because the reset to the nearest rotation matrix (18) locally reduces the distance to the true rotation, the SVD projection can be applied at every step without hurting convergence.

Reading between the lines

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

  • If the persistent-excitation integral (19) is monitored online, the same framework could serve as a fault detector: a drop in its smallest eigenvalue would warn that the current sensor set or maneuver no longer supports attitude estimation, before errors diverge.
  • For the Pitot and barometer/range modalities, the time-varying inertial vector $b_i(t)$ is under the vehicle's control through its trajectory; this suggests explicit maneuver-design rules could be derived to guarantee (19), something the paper leaves open.
  • The deterministic guarantee does not by itself quantify how measurement noise or model bias (wind, non-flat ground, unknown landmark heights) degrades the estimate; a natural next step would be an error-bound or bias-observability analysis for those cases.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the GES claim follows from an external LTV stability theorem applied to an exactly linearized state-space model, with no fitted parameter used as a prediction and no load-bearing self-citation chain.

full rationale

The derivation chain is self-contained algebra followed by an externally imported stability theorem. The authors define xB = vec(R^T), derive the exact LTV dynamics (10) and the exact linear output map (11)-(12), so (13) is not an approximation or ansatz that hides the target result. Lemma 1 rewrites the uniform observability Gramian condition using u_i(s) = b_i(s) ⊗ R(s)a_i(s); its proof in Appendix A is a direct computation of the transition matrix and Gramian, and the equivalence is algebraic rather than definitional smuggling. The GES guarantee for the Kalman filter (14) is taken from Hamel and Samson [20], an external published theorem, and the paper does not claim to prove or rename it; the cited theorem's assumptions (uniform observability of an LTV pair) do not include the conclusion that the attitude estimate converges. [21] is likewise an external elementary LTV result. The only self-citation with author overlap is [17], used contextually for the LTV framework and not as the load-bearing proof. There is no fitted-input-called-prediction pattern: M and Q are tuning matrices in the deterministic filter, are not estimated from data to force convergence, and do not enter the stability theorem. The Monte Carlo simulations are illustrative, not predictions derived from fitted parameters. One genuine limitation is candidly acknowledged in the text: the SVD projection and reset (18) preserve the geometric structure only 'at least locally,' and the Pythagorean orthogonality argument is only approximate, so the global theorem applies strictly to the reset-free continuous filter rather than to every element of Algorithm 1. That is a scoping gap between theorem and implementation, not a circularity, since the theorem itself does not reduce to its inputs. Overall, the central claim has independent content and no circular step can be exhibited.

Assumptions & free parameters 2 free parameters · 4 assumptions · 0 invented entities

The central theorem relies on a standard observability-implication theorem, exact angular velocity knowledge, idealized sensor models for the scalar modalities, and uniform positive definiteness of filter weights. No new physical entities are introduced.

free parameters (2)
  • Sensor noise covariances (gyroscope, accelerometer, magnetometer) = 0.001I3, 0.001I3, 0.01I3 (Table I)
    Chosen for the Monte Carlo simulation in Section VI. They do not enter the main theorem, but they are hand-selected simulation parameters.
  • Sensor sampling frequencies = 1000 Hz for IMU, 100 Hz for other sensors (Table I)
    Used in the discretization and simulation; implementation parameters, not fitted to reality.
assumptions (4)
  • standard math Uniform observability implies GES of the deterministic Kalman filter for LTV systems (Theorem 2.3 of [20])
    Invoked in Section IV to assert GES under condition (19).
  • domain assumption The angular velocity ω(t) is measured exactly (or with zero-mean noise) and the vectors a_i, b_i are known
    Required for state matrix A(t) and output matrix C(t); violated under gyro bias or unknown wind.
  • domain assumption The measurement model (3) captures the sensor modalities: negligible wind for Pitot, flat ground for range sensor, known landmark height difference
    Section III examples 2-4; if these idealized conditions fail, the scalar measurements contain bias.
  • domain assumption M(t) and Q(t) are uniformly positive definite
    Required for the Lyapunov argument in Section IV to guarantee negative definite V-dot.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Attitude Estimation Using Scalar Measurements." pith.science (2026). https://pith.science/paper/LQBMZVHT

@misc{pith2026250607068,
  author       = {Pith},
  title        = {Pith review of: Attitude Estimation Using Scalar Measurements},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/LQBMZVHT}},
  note         = {Machine review of arXiv:2506.07068}
}
read the original abstract

This paper revisits the problem of orientation estimation for rigid bodies through a novel framework based on scalar measurements. Unlike traditional vector-based methods, the proposed approach enables selective utilization of only the reliable axes of vector measurements while seamlessly incorporating alternative scalar modalities such as Pitot tubes, barometers with range sensors, and landmark-based constraints. The estimation problem is reformulated within a linear time-varying (LTV) framework, allowing the application of a deterministic linear Kalman filter. This design guarantees Global Uniform Exponential Stability (GES) under the Uniform Observability (UO) condition. Simulation results demonstrate the effectiveness of the proposed approach in achieving robust and accurate attitude estimation, even with partial vector measurements that simulate sensor axis failure.

Figures

Figures reproduced from arXiv: 2506.07068 by the authors.

Figure 1
Figure 1. Illustration of different sources of scalar measurements. IV. MAIN RESULTS In this section, we formulate the attitude kinematics as a state-space model, where the states correspond to the columns of the transposed rotation matrix expressed in the body frame. The output vector is derived from the generic measurement model (3), resulting in a linear time-varying (LTV) system. We then analyze the system’s uniform obser… view at source ↗
Figure 2
Figure 2. Illustration of the proposed attitude estimation approach [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Average attitude estimation errors. VII. CONCLUSION This work proposes a deterministic framework for attitude estimation using scalar measurements formulated within a linear time-varying (LTV) setting. The observer leverages a linear Kalman filter without linearization, resulting in a robust and simplified design. A closed-form matrix exponential is used in the discrete-time implementation to ensure accurate state t… view at source ↗

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

23 extracted references · 23 canonical work pages

  1. [1]

    Three-axis attitude determination from vector observations,

    M. D. Shuster and S. D. Oh, “Three-axis attitude determination from vector observations,” Journal of Guidance and Control , vol. 4, no. 1, pp. 70–77, 1981

    work page 1981

  2. [2]

    Attitude determination using vector observations and the singular value decomposition,

    F. L. Markley, “Attitude determination using vector observations and the singular value decomposition,” Journal of the Astronautical Sciences , vol. 36, no. 3, pp. 245–258, 1988

    work page 1988

  3. [3]

    Nonlinear complementary filters on the special orthogonal group,

    R. Mahony, T. Hamel, and J. M. Pflimlin, “Nonlinear complementary filters on the special orthogonal group,” IEEE Transactions on Automatic Control, vol. 53, no. 5, pp. 1203–1218, 2008

    work page 2008

  4. [4]

    Sensor-based globally asymp- totically stable filters for attitude estimation: Analysis, design, and performance evaluation,

    P. Batista, C. Silvestre, and P. Oliveira, “Sensor-based globally asymp- totically stable filters for attitude estimation: Analysis, design, and performance evaluation,” IEEE Transactions on Automatic Control , vol. 57, no. 8, pp. 2095–2100, 2012

    work page 2012

  5. [5]

    Nonlinear estimator design on the special orthogonal group using vector measurements directly,

    D. E. Zlotnik and J. R. Forbes, “Nonlinear estimator design on the special orthogonal group using vector measurements directly,” IEEE Transactions on Automatic Control , vol. 62, no. 1, pp. 149–160, 2016

    work page 2016

  6. [6]

    Attitude stabilization of a VTOL quadrotor aircraft,

    A. Tayebi and S. McGilvray, “Attitude stabilization of a VTOL quadrotor aircraft,” IEEE Transactions on Control Systems Technology , vol. 14, no. 3, pp. 562–571, 2006

    work page 2006

  7. [7]

    Rigid body attitude estimation based on the Lagrange–d’Alembert principle,

    M. Izadi and A. K. Sanyal, “Rigid body attitude estimation based on the Lagrange–d’Alembert principle,” Automatica, vol. 50, no. 10, pp. 2570–2577, 2014

    work page 2014

  8. [8]

    Survey of nonlinear atti- tude estimation methods,

    J. L. Crassidis, F. L. Markley, and Y . Cheng, “Survey of nonlinear atti- tude estimation methods,” Journal of Guidance, Control, and Dynamics, vol. 30, no. 1, pp. 12–28, 2007

    work page 2007

Show all 23 references
  1. [9]

    The invariant extended Kalman filter as a stable observer,

    A. Barrau and S. Bonnabel, “The invariant extended Kalman filter as a stable observer,” IEEE Transactions on Automatic Control , vol. 62, no. 4, pp. 1797–1812, 2017

    2017

  2. [10]

    Invariant Kalman filtering,

    ——, “Invariant Kalman filtering,” Annual Review of Control, Robotics, and Autonomous Systems , vol. 1, no. 1, pp. 237–257, 2018

    2018

  3. [11]

    Imple- mentation of a nonlinear attitude estimator for aerial robotic vehicles,

    M.-D. Hua, G. Ducard, T. Hamel, R. Mahony, and K. Rudin, “Imple- mentation of a nonlinear attitude estimator for aerial robotic vehicles,” IEEE Transactions on Control Systems Technology , vol. 22, no. 1, pp. 201–213, 2013

    2013

  4. [12]

    On the design of attitude complementary filters on so(3),

    S. Berkane and A. Tayebi, “On the design of attitude complementary filters on so(3),” IEEE Transactions on Automatic Control , vol. 63, no. 3, pp. 880–887, 2017

    2017

  5. [13]

    A nonlinear navigation observer using imu and generic position information,

    S. Berkane, A. Tayebi, and S. D. Marco, “A nonlinear navigation observer using imu and generic position information,” Automatica, vol. 127, p. 109513, 2021

    2021

  6. [14]

    Nonlinear observers design for vision-aided inertial navigation systems,

    M. Wang, S. Berkane, and A. Tayebi, “Nonlinear observers design for vision-aided inertial navigation systems,” IEEE Transactions on Automatic Control, vol. 67, no. 4, pp. 1853–1868, 2022

    2022

  7. [15]

    A GES attitude observer with single vector observations,

    P. Batista, C. Silvestre, and P. Oliveira, “A GES attitude observer with single vector observations,” Automatica, vol. 48, no. 2, pp. 388–395, 2012

    2012

  8. [16]

    Globally exponentially stable cascade observers for attitude estimation,

    ——, “Globally exponentially stable cascade observers for attitude estimation,” Control Engineering Practice, vol. 20, no. 2, pp. 148–155, 2012

    2012

  9. [17]

    A generic observer design for inertial navigation systems using an LTV framework,

    S. Benahmed and S. Berkane, “A generic observer design for inertial navigation systems using an LTV framework,” arXiv:2410.03846, 2025

    2025

  10. [18]

    Estimation of airspeed, angle of attack, and sideslip for small unmanned aerial vehicles (UA Vs) using a micro-pitot tube,

    G. Ariante, S. Ponte, U. Papa, and G. D. Core, “Estimation of airspeed, angle of attack, and sideslip for small unmanned aerial vehicles (UA Vs) using a micro-pitot tube,” Electronics, vol. 10, no. 19, p. 2325, 2021

    2021

  11. [19]

    Pitot tube measure-aided air velocity and attitude estimation in GNSS denied environment,

    T. L. de Oliveira, P. van Goor, T. Hamel, R. Mahony, and C. Sam- son, “Pitot tube measure-aided air velocity and attitude estimation in GNSS denied environment,” European Journal of Control, vol. 2024, p. 101070, 2024

    2024

  12. [20]

    Position estimation from direction or range measurements,

    T. Hamel and C. Samson, “Position estimation from direction or range measurements,” Automatica, vol. 82, pp. 137–144, 2017

    2017

  13. [21]

    Uniform exponential stability of linear time- varying systems: revisited,

    A. Lorıa and E. Panteley, “Uniform exponential stability of linear time- varying systems: revisited,” Systems & Control Letters , vol. 47, no. 1, pp. 13–24, 2002

    2002

  14. [22]

    Solution methods to the nearest rotation matrix problem in R3: A comparative survey,

    S. Sarabandi and F. Thomas, “Solution methods to the nearest rotation matrix problem in R3: A comparative survey,” Numerical Linear Alge- bra with Applications , vol. 30, no. 5, p. e2492, 2023

    2023

  15. [23]

    Analysis of non-linear attitude observers for time-varying reference measurements,

    J. Trumpf, R. Mahony, T. Hamel, and C. Lageman, “Analysis of non-linear attitude observers for time-varying reference measurements,” IEEE Transactions on Automatic Control , vol. 57, no. 11, pp. 2789– 2800, 2012

    2012

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.