pith. sign in

arxiv: 2509.04060 · v3 · submitted 2025-09-04 · 📡 eess.SY · cs.SY

Physics-Informed Detection of Friction Anomalies in Satellite Reaction Wheels

Alejandro Penacho Riveiros , Nicola Bastianello , Karl H. Johansson , Matthieu Barreau This is my paper

Pith reviewed 2026-05-18 19:38 UTC · model grok-4.3

classification 📡 eess.SY cs.SY
keywords friction anomaly detectionreaction wheel assembliessatellite health monitoringhybrid systems theorychangepoint detectionphysics-informed classificationanomaly detection
0
0 comments X

The pith

A combined model-based and data-driven algorithm detects friction anomalies in satellite reaction wheels with around 95% accuracy.

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

The paper develops an automated system to monitor friction in satellite reaction wheel assemblies and identify anomalies that need attention. It first applies a physics-based model from hybrid systems theory, incorporating changepoint detection, dynamic programming, and maximum likelihood, to pull out key information from the data. A classifier, trained mostly on nominal cases from a high-fidelity simulator, then labels the wheel status. Sympathetic readers would value this because the rapid increase in satellites makes manual monitoring impractical, and early anomaly detection prevents failures in critical space systems.

Core claim

The central discovery is that a physics-informed approach using hybrid systems theory to extract relevant features from reaction wheel friction data, followed by a classifier trained on high-fidelity simulated labelled data, can distinguish nominal operation from several anomaly types with approximately 95 percent accuracy.

What carries the argument

The hybrid systems model for feature extraction that integrates changepoint detection, dynamic programming, and maximum likelihood to prepare data for classification.

If this is right

  • The algorithm enables automated detection of friction anomalies requiring preventive measures.
  • It reduces the human workload for monitoring the growing number of satellites in orbit.
  • Training on mostly nominal simulated data still yields satisfactory classification performance.
  • Integration of model-based extraction with data-based classification improves reliability for on-board status determination.

Where Pith is reading between the lines

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

  • This method could extend to anomaly detection in other satellite components with similar data characteristics.
  • Real-world validation on actual satellite telemetry would confirm if the simulator captures anomalies accurately enough.
  • Similar hybrid techniques might apply to fault detection in other engineering systems like aircraft or vehicles.
  • The approach suggests that limited anomalous data can still support effective classifiers when paired with strong physical models.

Load-bearing premise

The labelled data from the high-fidelity simulator must accurately reflect the real friction anomalies that occur on actual satellites.

What would settle it

Running the algorithm on real satellite reaction wheel telemetry data and checking its classifications against verified anomaly occurrences or expert manual analysis would test if the reported accuracy holds outside simulation.

Figures

Figures reproduced from arXiv: 2509.04060 by Alejandro Penacho Riveiros, Karl H. Johansson, Matthieu Barreau, Nicola Bastianello.

Figure 1
Figure 1. Figure 1: Cross-section of a Reaction Wheel Assembly, adapted [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Examples of runs labeled with different anomalies. F [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 4
Figure 4. Figure 4: Visualization of the two friction systems described [PITH_FULL_IMAGE:figures/full_fig_p005_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Diagram of the complete algorithm presented. The not [PITH_FULL_IMAGE:figures/full_fig_p006_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Example of possible transitions during the dynamic p [PITH_FULL_IMAGE:figures/full_fig_p009_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Left: value of ARL0.1% for different window sizes w and bias Wb, assuming changes in friction of ∆fcp = 3σv. Right: effect of window size and friction change on ARL0.1%, with Wb = 10−4 . As a last note, the average run lengths obtained in these analyses assume that the data contains no abnormalities, that is, the friction is just a combination of dry and viscous friction and Gaussian noise as given by (4a)… view at source ↗
Figure 9
Figure 9. Figure 9: Survival function of the absolute error of the fricti [PITH_FULL_IMAGE:figures/full_fig_p012_9.png] view at source ↗
Figure 8
Figure 8. Figure 8: Excess RMSE in the datapoints of the TAS dataset. The t [PITH_FULL_IMAGE:figures/full_fig_p012_8.png] view at source ↗
Figure 10
Figure 10. Figure 10: Relation between the number of iterations run by the [PITH_FULL_IMAGE:figures/full_fig_p013_10.png] view at source ↗
Figure 11
Figure 11. Figure 11: Anomaly detection probabilities obtained with 30 r [PITH_FULL_IMAGE:figures/full_fig_p013_11.png] view at source ↗
Figure 13
Figure 13. Figure 13: Distribution of computation time taken by each step [PITH_FULL_IMAGE:figures/full_fig_p014_13.png] view at source ↗
Figure 12
Figure 12. Figure 12: Train and validation loss with varying number of bin [PITH_FULL_IMAGE:figures/full_fig_p014_12.png] view at source ↗
read the original abstract

As the number of satellites in orbit has increased exponentially in recent years, ensuring their correct functionality has started to require automated methods to decrease human workload. In this work, we present an algorithm that analyzes the on-board data related to friction from the Reaction Wheel Assemblies (RWA) of a satellite and determines their operating status, distinguishing between nominal status and several possible anomalies that require preventive measures to be taken. The algorithm first uses a model based on hybrid systems theory to extract the information relevant to the problem. The extraction process combines techniques in changepoint detection, dynamic programming, and maximum likelihood in a structured way. A classifier then uses the extracted information to determine the status of the RWA. This last classifier has been previously trained with a labelled dataset produced by a high-fidelity simulator, comprised for the most part of nominal data. The final algorithm combines model-based and data-based approaches to obtain satisfactory results with an accuracy around 95%.

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

2 major / 2 minor

Summary. The paper proposes a hybrid algorithm for detecting friction anomalies in satellite reaction wheel assemblies (RWAs). It first applies a model-based approach grounded in hybrid systems theory, using changepoint detection, dynamic programming, and maximum likelihood estimation to extract relevant features from on-board friction data. These features then feed a classifier trained on a labelled dataset generated by a high-fidelity simulator (mostly nominal cases with anomalies introduced via parameter changes). The combined model-based and data-driven method is reported to achieve approximately 95% accuracy in distinguishing nominal operation from several anomaly types.

Significance. If the simulator faithfully reproduces the statistical signatures of real orbital friction anomalies, the work offers a practical physics-informed pipeline that reduces reliance on manual telemetry review for the growing satellite fleet. The explicit integration of hybrid-systems feature extraction with supervised classification is a constructive strength; it avoids purely black-box methods while still leveraging data-driven performance. Reproducible simulator-based evaluation and the structured extraction pipeline are positive elements that could support follow-on flight validation.

major comments (2)
  1. [Abstract / Results] Abstract and results section: The headline claim of ~95% accuracy is obtained exclusively on held-out simulator samples. No quantitative comparison to real satellite telemetry is presented, nor are error bars, cross-validation details, or sensitivity to post-hoc feature-extraction choices reported. Because the central performance guarantee rests on the simulator-to-reality transfer, this omission is load-bearing for any claim of operational utility.
  2. [Methods / Simulator description] Methods: The high-fidelity simulator is used to generate labelled anomalies by altering friction parameters, yet the manuscript does not demonstrate that the resulting torque profiles and changepoint statistics match those observed in actual orbital degradation (e.g., temperature-dependent viscosity or microgravity preload effects). Without such a statistical equivalence test, the downstream classifier’s 95% figure cannot be extrapolated beyond the simulator.
minor comments (2)
  1. [Methods] Notation for the hybrid-system modes and the dynamic-programming cost function should be introduced with explicit equation numbers to improve traceability from feature extraction to classifier input.
  2. [Figures] Figure captions for the changepoint detection examples would benefit from explicit indication of which segments correspond to nominal versus anomalous regimes.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive and detailed review. The comments highlight important aspects of simulator validation and performance reporting that we address below. We have revised the manuscript to incorporate additional details, metrics, and discussion while maintaining the core contributions of the hybrid physics-informed approach.

read point-by-point responses

  1. Referee: [Abstract / Results] Abstract and results section: The headline claim of ~95% accuracy is obtained exclusively on held-out simulator samples. No quantitative comparison to real satellite telemetry is presented, nor are error bars, cross-validation details, or sensitivity to post-hoc feature-extraction choices reported. Because the central performance guarantee rests on the simulator-to-reality transfer, this omission is load-bearing for any claim of operational utility.

    Authors: We acknowledge that the reported accuracy is evaluated on held-out simulator data and that direct quantitative comparison to real telemetry would be valuable for operational claims. Labeled real anomaly data remains scarce due to the rarity of events and proprietary constraints on satellite telemetry. The simulator is constructed from hybrid systems models and physical friction parameters drawn from established RWA literature. In the revised version we add: (i) error bars from repeated simulation trials, (ii) explicit description of the stratified cross-validation procedure, and (iii) sensitivity analysis with respect to changepoint detection hyperparameters. We also insert a limitations subsection that explicitly discusses the simulator-to-reality gap and the desirability of future on-orbit validation. revision: partial

  2. Referee: [Methods / Simulator description] Methods: The high-fidelity simulator is used to generate labelled anomalies by altering friction parameters, yet the manuscript does not demonstrate that the resulting torque profiles and changepoint statistics match those observed in actual orbital degradation (e.g., temperature-dependent viscosity or microgravity preload effects). Without such a statistical equivalence test, the downstream classifier’s 95% figure cannot be extrapolated beyond the simulator.

    Authors: The simulator implements physics-based friction models that include temperature-dependent viscosity and preload effects calibrated against published RWA degradation studies. Anomaly cases are generated by controlled parameter perturbations chosen to produce torque signatures consistent with known failure modes. A direct statistical equivalence test against real orbital telemetry is not feasible within the present study because sufficiently labeled real anomaly datasets are not publicly available. In revision we expand the simulator section with explicit justification of each parameter range, add quantitative metrics comparing simulated versus expected changepoint statistics under nominal conditions, and include a discussion of remaining modeling assumptions. revision: partial

Circularity Check

0 steps flagged

No significant circularity; derivation relies on independent simulator data and standard statistical methods

full rationale

The paper's core pipeline extracts features from reaction-wheel friction signals using hybrid-systems changepoint detection, dynamic programming, and maximum-likelihood estimation, then feeds those features into a classifier trained on a separate high-fidelity simulator dataset. The reported 95% accuracy is obtained on held-out simulator samples rather than on the training set itself, so the performance metric is not forced by construction. No self-definitional loops, fitted parameters renamed as predictions, or load-bearing self-citations that collapse the central claim appear in the described methodology. The approach therefore remains self-contained against the external simulator benchmark.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

The central claim rests on the assumption that the high-fidelity simulator faithfully reproduces real friction dynamics and anomaly signatures. No free parameters, axioms, or invented entities are explicitly listed in the abstract.

pith-pipeline@v0.9.0 · 5702 in / 1090 out tokens · 28395 ms · 2026-05-18T19:38:29.071860+00:00 · methodology

discussion (0)

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

Reference graph

Works this paper leans on

47 extracted references · 47 canonical work pages

  1. [1]

    General Catalog of Artificial Space Obje cts,

    J. C. McDowell, “General Catalog of Artificial Space Obje cts,” 2023

    work page 2023

  2. [2]

    STS-61 Space Shuttle Mission Report,

    R. W. Fricke, “STS-61 Space Shuttle Mission Report,” NAS A, Lyndon B. Johnson Space Center, Houston, Texas, Tech. Rep., 1994

    work page 1994

  3. [3]

    Generic Model of a Satellite Attitude Control System,

    J. Narkiewicz, M. Sochacki, and B. Zakrzewski, “Generic Model of a Satellite Attitude Control System,” International Journal of Aerospace Engineering, vol. 2020, p. e5352019, 2020

    work page 2020

  4. [4]

    Curing XMM-Newton’s reaction wheel cage instability: the in-flight re-lubrication exper ience,

    M. Pantaleoni, P . Chapman, R. Harris, M. G. Kirsch, R. Kre sken, J. Martin, P . McMahon, A. Mcdonald, F. Schmidt, T. Strandber g, D. Webert, and U. Weissmann, “Curing XMM-Newton’s reaction wheel cage instability: the in-flight re-lubrication exper ience,” in SpaceOps 2014 Conference , ser. SpaceOps Conferences. American Institute of Aeronautics and Astron...

    work page doi:10.2514/6.2014-1875 2014

  5. [5]

    In-Flight Position Calibration of the Cassin i Articulated Reaction Wheel Assembly,

    T. Brown, “In-Flight Position Calibration of the Cassin i Articulated Reaction Wheel Assembly,” in AIAA Guidance, Navigation, and Control Conference, ser. Guidance, Navigation, and Control and Co-located Conferences. American Institute of Aeronautics and Astron autics, 2012

    work page 2012

  6. [6]

    Cage Instability of XMM-Newton’s Reaction W heels Dis- covered during the Development of an Early Degradation Warn ing System,

    M. Kirsch, “Cage Instability of XMM-Newton’s Reaction W heels Dis- covered during the Development of an Early Degradation Warn ing System,” in SpaceOps 2012 Conference , ser. SpaceOps Conferences. American Institute of Aeronautics and Astronautics, 2012

    work page 2012

  7. [7]

    Kepler Mission Operations Response to Wheel Anomalies,

    K. A. Larson, K. M. McCalmont, C. A. Peterson, and S. E. Ros s, “Kepler Mission Operations Response to Wheel Anomalies,” in SpaceOps 2014 Conference. American Institute of Aeronautics and Astronautics, 2014 . [Online]. Available: https://arc.aiaa.org/doi/abs/10. 2514/6.2014-1882

    work page 2014

  8. [8]

    Dawn’s exploration of V esta,

    M. D. Rayman and R. A. Mase, “Dawn’s exploration of V esta, ” Acta Astronautica, vol. 94, no. 1, pp. 159–167, 2014

    work page 2014

  9. [9]

    Solar Dynamics Observatory Reaction Whee l Bearing Friction Increase: Detection, Analysis, and Impacts,

    F. M. Ekinci, “Solar Dynamics Observatory Reaction Whee l Bearing Friction Increase: Detection, Analysis, and Impacts,” American Institute of Aeronautics and Astronautics , 2014

    work page 2014

  10. [10]

    Operations w ith the new FUSE observatory: three-axis control with one react ion wheel,

    D. Sahnow, J. Kruk, T. Ake, B.-G. Andersson, A. Berman, W . Blair, R. Boyer, J. Caplinger, H. Calvani, T. Civeit, W. V an, W. Dixo n, M. England, M. Kaiser, H. Moos, and B. Roberts, “Operations w ith the new FUSE observatory: three-axis control with one react ion wheel,” Proceedings of SPIE - The International Society for Optical Engineer- ing, vol. 6266, 2006

    work page 2006

  11. [11]

    A NEWL Y DISCOVERED BRANCH O F THE FAULT TREE EXPLAINING SYSTEMIC REACTION WHEEL FAILURES AND ANOMALIES,

    W. E. Bialke and E. Hansell, “A NEWL Y DISCOVERED BRANCH O F THE FAULT TREE EXPLAINING SYSTEMIC REACTION WHEEL FAILURES AND ANOMALIES,” in ESMATS 2017, Hatfield, 2017

    work page 2017

  12. [12]

    Reaction Wheel Performance Characterization Using the Kepler Space craft as a Case Study,

    J. Kampmeier, R. Larsen, L. F. Migliorini, and K. A. Lars on, “Reaction Wheel Performance Characterization Using the Kepler Space craft as a Case Study,” in 2018 SpaceOps Conference, ser. SpaceOps Conferences. American Institute of Aeronautics and Astronautics, 2018

    work page 2018

  13. [13]

    Improving Spacecraft Health Monitoring with A utomatic Anomaly Detection Techniques,

    S. Fuertes, G. Picart, J.-Y . Tourneret, L. Chaari, A. Fe rrari, and C. Richard, “Improving Spacecraft Health Monitoring with A utomatic Anomaly Detection Techniques,” in SpaceOps 2016 Conference , ser. SpaceOps Conferences. American Institute of Aeronautics a nd Astro- nautics, 2016

    work page 2016

  14. [14]

    An Anomaly Detec tion Method for Spacecraft Using Relevance V ector Learning,

    R. Fujimaki, T. Y airi, and K. Machida, “An Anomaly Detec tion Method for Spacecraft Using Relevance V ector Learning,” in Advances in Knowledge Discovery and Data Mining , ser. Lecture Notes in Computer Science. Berlin, Heidelberg: Springer, 2005

    work page 2005

  15. [15]

    Fault detection and d iagnosis for spacecraft using principal component analysis and supp ort vector machines,

    Y . Gao, T. Y ang, N. Xing, and M. Xu, “Fault detection and d iagnosis for spacecraft using principal component analysis and supp ort vector machines,” in 2012 7th IEEE Conference on Industrial Electronics and Applications (ICIEA) , 2012, pp. 1984–1988

    work page 2012

  16. [16]

    Machine Learning Methods for Spacecraft Telemetry Mining,

    S. K. Ibrahim, A. Ahmed, M. A. E. Zeidan, and I. E. Ziedan, “Machine Learning Methods for Spacecraft Telemetry Mining,” IEEE Transactions on Aerospace and Electronic Systems , vol. 55, no. 4, 2019

    work page 2019

  17. [17]

    Mining distance-based ou tliers in near linear time with randomization and a simple pruning rul e,

    S. D. Bay and M. Schwabacher, “Mining distance-based ou tliers in near linear time with randomization and a simple pruning rul e,” in Proceedings of the ninth ACM SIGKDD international conferen ce on Knowledge discovery and data mining , ser. KDD ’03. New Y ork, NY , USA: Association for Computing Machinery, 2003, pp. 29–38

    work page 2003

  18. [18]

    Inductive System Health Monitoring,

    D. L. Iverson, “Inductive System Health Monitoring,” i n ResearchGate, CSREA, Las V egas, Nevada, 2004

    work page 2004

  19. [19]

    Develop ment and V alidation of Reaction Wheel Disturbance Models: Empirica l Model,

    R. A. Masterson, D. W. Miller, and R. L. Grogan, “Develop ment and V alidation of Reaction Wheel Disturbance Models: Empirica l Model,” Journal of Sound and Vibration , vol. 249, no. 3, pp. 575–598, 2002

    work page 2002

  20. [20]

    M icrovibration simulation of reaction wheel ball bearings,

    M. M. Longato, T. Hughes, V . Y otov, and G. S. Aglietti, “M icrovibration simulation of reaction wheel ball bearings,” Journal of Sound and Vibration, vol. 567, p. 117909, 2023

    work page 2023

  21. [21]

    Parameter Estimation for Fault Diagnosis in Nonlinear Systems by ANFI S,

    B. Bellali, A. Hazzab, I. K. Bousserhane, and D. Lefebvr e, “Parameter Estimation for Fault Diagnosis in Nonlinear Systems by ANFI S,” Procedia Engineering, vol. 29, pp. 2016–2021, 2012

    work page 2016

  22. [22]

    Efficient machine learning based techniques for fault dete ction and identification in spacecraft reaction wheel,

    T. S. Abdel Aziz, G. I. Salama, M. S. Mohamed, and S. Husse in, “Efficient machine learning based techniques for fault dete ction and identification in spacecraft reaction wheel,” Aerospace Systems , vol. 7, no. 4, pp. 815–828, 2024

    work page 2024

  23. [23]

    Using Machine L earning to Automatically Detect Anomalous Spacecraft Behavior from T elemetry Data,

    K. Naik, A. Holmgren, and J. Kenworthy, “Using Machine L earning to Automatically Detect Anomalous Spacecraft Behavior from T elemetry Data,” in 2020 IEEE Aerospace Conference , 2020, pp. 1–14

    work page 2020

  24. [24]

    Fault detection and diagnosis of rotating machinery,

    K. Loparo, M. Adams, W. Lin, M. Abdel-Magied, and N. Afsh ari, “Fault detection and diagnosis of rotating machinery,” IEEE Transactions on Industrial Electronics, vol. 47, no. 5, pp. 1005–1014, 2000

    work page 2000

  25. [25]

    A compr ehensive review of artificial intelligence-based approaches for rol ling element bearing PHM: shallow and deep learning,

    M. Hamadache, J. H. Jung, J. Park, and B. D. Y oun, “A compr ehensive review of artificial intelligence-based approaches for rol ling element bearing PHM: shallow and deep learning,” JMST Advances, vol. 1, no. 1, pp. 125–151, 2019

    work page 2019

  26. [26]

    Fault detection and diag nosis in low speed rolling element bearings Part II: The use of nearest ne ighbour classification,

    C. K. Mechefske and J. Mathew, “Fault detection and diag nosis in low speed rolling element bearings Part II: The use of nearest ne ighbour classification,” Mechanical Systems and Signal Processing , vol. 6, no. 4, pp. 309–316, 1992

    work page 1992

  27. [27]

    Detection of common mot or bearing faults using frequency-domain vibration signals a nd a neural network based approach,

    B. Li, G. Goddu, and M.-Y . Chow, “Detection of common mot or bearing faults using frequency-domain vibration signals a nd a neural network based approach,” in Proceedings of the 1998 American Control Conference. ACC, vol. 4, 1998, pp. 2032–2036 vol.4

    work page 1998

  28. [28]

    Effect of number o f features on classification of roller bearing faults using SVM and PSVM ,

    V . Sugumaran and K. I. Ramachandran, “Effect of number o f features on classification of roller bearing faults using SVM and PSVM ,” Expert Systems with Applications , vol. 38, no. 4, pp. 4088–4096, 2011

    work page 2011

  29. [29]

    Fault prognosis f or rotating electrical machines monitoring using recursive least squa re,

    M. Rocchi, F. Mosciaro, F. Grottesi, M. Scortichini, A. Giantomassi, M. Pirro, M. Grisostomi, and G. Ippoliti, “Fault prognosis f or rotating electrical machines monitoring using recursive least squa re,” in 2014 6th European Embedded Design in Education and Research Conf erence (EDERC), 2014, pp. 269–273

    work page 2014

  30. [30]

    Fau lt Diagnosis of Rotating Machinery With Limited Expert Intera ction: A Multicriteria Active Learning Approach Based on Broad Lea rning System,

    Z. Liu, J. Zhang, X. He, Q. Zhang, G. Sun, and D. Zhou, “Fau lt Diagnosis of Rotating Machinery With Limited Expert Intera ction: A Multicriteria Active Learning Approach Based on Broad Lea rning System,” IEEE Transactions on Control Systems Technology , vol. 31, no. 2, pp. 953–960, 2023

    work page 2023

  31. [31]

    V alidation of a physics-based model of a rotating machine for synthetic data generation in hybrid diagnosis,

    U. Leturiondo, O. Salgado, and D. Galar, “V alidation of a physics-based model of a rotating machine for synthetic data generation in hybrid diagnosis,” Structural Health Monitoring , vol. 16, no. 4, 2017

    work page 2017

  32. [32]

    Ma chine Fault Classification using Hamiltonian Neural Networks,

    J. Shen, J. Chowdhury, S. Banerjee, and G. Terejanu, “Ma chine Fault Classification using Hamiltonian Neural Networks,” 2023

    work page 2023

  33. [33]

    Low V elocity Fricti on Compen- sation and Feedforward Solution based on Repetitive Contro l,

    E. Tung, G. Anwar, and M. Tomizuka, “Low V elocity Fricti on Compen- sation and Feedforward Solution based on Repetitive Contro l,” in 1991 American Control Conference , 2003

    work page 1991

  34. [34]

    Experimental study on low velocity friction compensation and tracking control,

    M. Popovic, G. Liu, and A. A. Goldenberg, “Experimental study on low velocity friction compensation and tracking control,” Journal of Automatic Control, vol. 13, 2003

    work page 2003

  35. [35]

    Switching Mode Disturbance Observer for Fric tion Compensation in Linear Motors,

    M. Geissmann, H. P . Willi, W. Fischer, P . Kontopulos, an d T. J. Besselmann, “Switching Mode Disturbance Observer for Fric tion Compensation in Linear Motors,” IEEE Transactions on Control Systems Technology, vol. 32, no. 3, pp. 1040–1047, 2024

    work page 2024

  36. [36]

    A switching Gaussian proce ss latent force model for the identification of mechanical systems with a dis continuous nonlinearity,

    L. Marino and A. Cicirello, “A switching Gaussian proce ss latent force model for the identification of mechanical systems with a dis continuous nonlinearity,” Data-Centric Engineering , vol. 4, p. e18, 2023

    work page 2023

  37. [37]

    Armstrong-H´ elouvry, Control of Machines with Friction

    B. Armstrong-H´ elouvry, Control of Machines with Friction . Boston, MA: Springer US, 1991

    work page 1991

  38. [38]

    Slip-based tire-road friction estima tion,

    F. Gustafsson, “Slip-based tire-road friction estima tion,” Automatica, vol. 33, no. 6, pp. 1087–1099, 1997

    work page 1997

  39. [39]

    Estimating Friction Paramet ers in Reaction Wheels for Attitude Control,

    V . Carrara and H. K. Kuga, “Estimating Friction Paramet ers in Reaction Wheels for Attitude Control,” Mathematical Problems in Engineering , vol. 2013, no. 1, p. 249674, 2013

    work page 2013

  40. [40]

    Reaction Wheel Fric tion Analysis Methodology and On-orbit Experience,

    J. M. Hacker, J. Ying, and P . C. Lai, “Reaction Wheel Fric tion Analysis Methodology and On-orbit Experience,” in AIAA/AAS Astrodynamics Specialist Conference . American Institute of Aeronautics and Astro- nautics, 2023

    work page 2023

  41. [41]

    In-Flight Performance of Cassi ni Reaction Wheel Bearing Drag in 1997–2013,

    A. Y . Lee and E. K. Wang, “In-Flight Performance of Cassi ni Reaction Wheel Bearing Drag in 1997–2013,” Journal of Spacecraft and Rockets , vol. 52, no. 2, pp. 470–480, 2015

    work page 1997

  42. [42]

    Viscosity–temperature correlation for liquids,

    C. J. Seeton, “Viscosity–temperature correlation for liquids,” Tribology Letters, vol. 22, no. 1, pp. 67–78, 2006. JOURNAL OF LATEX CLASS FILES, VOL. 14, NO. 8, AUGUST 2021 16

    work page 2006

  43. [43]

    A tutorial on hidden Markov models and sele cted applications in speech recognition,

    L. Rabiner, “A tutorial on hidden Markov models and sele cted applications in speech recognition,” Proceedings of the IEEE , vol. 77, no. 2, pp. 257–286, 1989

    work page 1989

  44. [44]

    Gustafsson, Adaptive Filtering and Change Detection , 1st ed

    F. Gustafsson, Adaptive Filtering and Change Detection , 1st ed. Chich- ester New Y ork: Wiley, 2000

    work page 2000

  45. [45]

    The Large-Sample Distribution of the Like lihood Ratio for Testing Composite Hypotheses,

    S. S. Wilks, “The Large-Sample Distribution of the Like lihood Ratio for Testing Composite Hypotheses,” The Annals of Mathematical Statistics, vol. 9, no. 1, pp. 60–62, 1938

    work page 1938

  46. [46]

    D. P . Bertsekas, Dynamic Programming and Optimal Control . Nashua, NH: Athena Scientific, 2012

    work page 2012

  47. [47]

    Hastie, R

    T. Hastie, R. Tibshirani, and J. Friedman, The Elements of Statistical Learning, ser. Springer Series in Statistics. Springer, 2009. APPENDIX A BIOGRAPHY SECTION

    work page 2009