pith. sign in

arxiv: 2605.15851 · v1 · pith:CKPKAN7Wnew · submitted 2026-05-15 · 📡 eess.SY · cs.SY

Uncertainty Propagation under Residual Disturbances: A Smart-Home Case Study

Guanru Pan , Dirk Reinhardt , Sebastien Gros , Timm Faulwasser This is my paper

Pith reviewed 2026-05-20 16:42 UTC · model grok-4.3

classification 📡 eess.SY cs.SY
keywords uncertainty propagationresidual disturbancesdata-driven predictionpolynomial chaos expansionsdistributional consistencysmart homecausal predictors
0
0 comments X

The pith

Consolidating unmodeled disturbances into one residual term produces a causal stochastic predictor that supports efficient uncertainty quantification.

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

The paper introduces a data-driven approach that gathers all statistically unmodeled disturbances into a single residual quantity estimated from measurements. Under mild assumptions this residual yields a stochastic predictor that stays causal and reproduces the observed data distribution. The construction then permits direct application of polynomial chaos expansions and higher-order Chebyshev inequalities to compute uncertainty bounds without separate models for each disturbance source. The framework is demonstrated on experimental data collected from a smart home in Norway.

Core claim

Under mild assumptions, consolidating all unstructured disturbances into a single residual disturbance that can be estimated from data produces a stochastic predictor that is both causal and distributionally consistent, which in turn enables efficient uncertainty quantification through polynomial chaos expansions and higher-order Chebyshev inequalities.

What carries the argument

The residual disturbance, a single data-estimated quantity that aggregates all unstructured disturbances and defines the stochastic predictor.

If this is right

  • Uncertainty bounds can be computed with polynomial chaos expansions rather than Monte-Carlo sampling.
  • Higher-order Chebyshev inequalities supply rigorous, non-conservative probability bounds on prediction errors.
  • The same residual construction applies to any system where individual disturbances cannot be measured or modeled separately.
  • The method requires only input-output data and does not need a full statistical model of every disturbance.

Where Pith is reading between the lines

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

  • The residual approach could be combined with existing model-predictive controllers to improve robustness without explicit disturbance modeling.
  • Sensitivity of the consistency property to the choice of residual estimator remains open for further analysis.
  • The same consolidation idea might apply to networked systems where only aggregate measurements are available.

Load-bearing premise

All unstructured disturbances can be consolidated into one residual quantity that is estimable from data while keeping the resulting predictor causal and distributionally consistent.

What would settle it

A set of smart-home trajectories in which the polynomial-chaos predictions systematically deviate from the empirical distribution of measured outputs would show that the residual-based predictor is not distributionally consistent.

Figures

Figures reproduced from arXiv: 2605.15851 by Dirk Reinhardt, Guanru Pan, Sebastien Gros, Timm Faulwasser.

Figure 1
Figure 1. Figure 1: Conceptual block diagram of the residential build [PITH_FULL_IMAGE:figures/full_fig_p005_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Comparison of true and predicted thermal re [PITH_FULL_IMAGE:figures/full_fig_p006_2.png] view at source ↗
read the original abstract

This paper presents a data-driven framework for uncertainty propagation under unmeasured or statistically unmodeled (unstructured) disturbances. We consider residual disturbances, which consolidate all unstructured disturbances into a single quantity that can be estimated from data. Under mild assumptions, the resulting stochastic predictor is causal and distributionally consistent, enabling efficient uncertainty quantification through polynomial chaos expansions and higher-order Chebyshev inequalities. The proposed method is validated using experimental data from a smart home in Norway.

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 presents a data-driven framework for uncertainty propagation under unmeasured or statistically unmodeled disturbances by consolidating them into residual disturbances that are estimated from data. Under mild assumptions, the resulting stochastic predictor is claimed to be causal and distributionally consistent. This enables efficient uncertainty quantification via polynomial chaos expansions and higher-order Chebyshev inequalities. The method is validated on experimental data from a smart home in Norway.

Significance. If the residual estimation procedure can be shown to preserve causality (via strictly past data) and distributional consistency (including higher-order moments), the framework could offer a practical, computationally efficient approach to uncertainty quantification in systems with significant unstructured disturbances, such as smart-home energy or control applications. The experimental validation on real traces is a positive element, but the absence of detailed quantitative results, error analysis, or explicit assumption verification limits the assessed impact.

major comments (2)
  1. [Abstract and §3] Abstract and §3: The central claim that the stochastic predictor is causal and distributionally consistent under mild assumptions is load-bearing, yet the manuscript provides no explicit statement of the assumptions, no derivation showing that residual estimation uses only past data, and no proof that lumping preserves cross-correlations or higher-order moments needed for PCE and Chebyshev bounds.
  2. [§4] §4 (smart-home case study): The experimental validation must demonstrate on the concrete traces that the residual estimation procedure is strictly causal and that the lumped residual yields distributionally consistent predictions; without quantitative metrics (e.g., moment errors, bound tightness, or comparison to ground-truth uncertainty), the support for the claims remains insufficient.
minor comments (2)
  1. Clarify the precise definition and estimation algorithm for the residual disturbance (e.g., filtering or regression steps) with pseudocode or explicit equations.
  2. Add a table or figure summarizing quantitative validation results (prediction accuracy, uncertainty bound coverage) to strengthen the case-study section.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive comments on our manuscript. The concerns about explicit assumptions, derivations for causality and distributional consistency in §3, and the need for quantitative metrics in the smart-home case study are valid points that we will address through revisions. We respond to each major comment below.

read point-by-point responses

  1. Referee: [Abstract and §3] Abstract and §3: The central claim that the stochastic predictor is causal and distributionally consistent under mild assumptions is load-bearing, yet the manuscript provides no explicit statement of the assumptions, no derivation showing that residual estimation uses only past data, and no proof that lumping preserves cross-correlations or higher-order moments needed for PCE and Chebyshev bounds.

    Authors: We agree that the assumptions and supporting derivations should be stated more explicitly. In the revised manuscript we will add a dedicated subsection at the start of §3 that lists the mild assumptions in numbered form (additive unstructured disturbances, finite moments of order up to 2k, and stationarity of the residual process). We will also insert a short derivation showing that the residual estimation filter is strictly causal, using only observations up to time t-1. Regarding preservation of cross-correlations and higher-order moments, we will add a lemma proving that the lumped residual matches the empirical joint distribution of the total disturbance by construction of the data-driven estimator; this is sufficient to justify the subsequent PCE and Chebyshev steps. These additions will be placed before the main theorems. revision: yes

  2. Referee: [§4] §4 (smart-home case study): The experimental validation must demonstrate on the concrete traces that the residual estimation procedure is strictly causal and that the lumped residual yields distributionally consistent predictions; without quantitative metrics (e.g., moment errors, bound tightness, or comparison to ground-truth uncertainty), the support for the claims remains insufficient.

    Authors: We accept that the current experimental section lacks sufficient quantitative evidence. In the revision we will expand §4 with new tables reporting (i) the L2 error between estimated residual moments and sample moments computed directly from the Norwegian traces, (ii) the ratio of Chebyshev bound width to observed prediction interval width on held-out segments, and (iii) a side-by-side comparison of uncertainty envelopes with and without the residual term. We will also add a paragraph and accompanying plot that explicitly verifies causality by showing the residual at time t is formed exclusively from data up to t-1. These metrics will be computed on the same experimental traces already presented. revision: yes

Circularity Check

0 steps flagged

No significant circularity; derivation relies on independent assumptions and external data validation

full rationale

The paper introduces residual disturbances as a lumped quantity estimated from data and states that under mild assumptions the resulting stochastic predictor is causal and distributionally consistent. These assumptions are presented as external conditions rather than derived from or fitted to the target quantities; the framework then applies PCE and Chebyshev inequalities for UQ. Validation occurs on independent experimental traces from a smart-home system in Norway. No equations or steps are shown to reduce by construction to fitted parameters, self-citations, or renamed known results. The central claims therefore remain self-contained against the stated assumptions and the separate experimental check.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 1 invented entities

The framework rests on the concept of residual disturbances as a consolidation of unstructured effects and on mild assumptions for causality and consistency; no specific numerical free parameters or entities with independent evidence are detailed.

axioms (1)
  • domain assumption Mild assumptions ensure the stochastic predictor is causal and distributionally consistent when using residual disturbances estimated from data
    Invoked directly in the abstract as the foundation for the method's properties and efficiency claims.
invented entities (1)
  • residual disturbances no independent evidence
    purpose: Consolidate all unmeasured or statistically unmodeled disturbances into a single quantity estimable from data
    Introduced as the central modeling device to enable the data-driven stochastic predictor.

pith-pipeline@v0.9.0 · 5600 in / 1379 out tokens · 101988 ms · 2026-05-20T16:42:00.866099+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

255 extracted references · 255 canonical work pages

  1. [1]

    On a stochastic fundamental lemma and its use for data-driven optimal control , year =

    Pan, Guanru and Ou, Ruchuan and Faulwasser, Timm , journal =. On a stochastic fundamental lemma and its use for data-driven optimal control , year =

    work page

  2. [2]

    On data-driven stochastic output-feedback predictive control , year =

    Pan, Guanru and Ou, Ruchuan and Faulwasser, Timm , journal =. On data-driven stochastic output-feedback predictive control , year =

    work page

  3. [3]

    Towards data-driven stochastic predictive control , year =

    Pan, Guanru and Ou, Ruchuan and Faulwasser, Timm , journal =. Towards data-driven stochastic predictive control , year =

    work page

  4. [4]

    Distributionally robust uncertainty quantification via data-driven stochastic optimal control , year =

    Pan, Guanru and Faulwasser, Timm , journal =. Distributionally robust uncertainty quantification via data-driven stochastic optimal control , year =

    work page

  5. [5]

    Data-driven multiple shooting for stochastic optimal control , year =

    Ou, Ruchuan and Pan, Guanru and Faulwasser, Timm , journal =. Data-driven multiple shooting for stochastic optimal control , year =

    work page

  6. [6]

    Toward data-driven predictive control of multi-energy distribution systems , year =

    Bilgic, Deborah and Koch, Alexander and Pan, Guanru and Faulwasser, Timm , journal =. Toward data-driven predictive control of multi-energy distribution systems , year =

    work page

  7. [7]

    IFAC-Pap

    Pan, Guanru and Stomberg, G. IFAC-Pap. , title =. 2021 , number =

    work page 2021

  8. [8]

    2023 , pages =

    Pan, Guanru and Faulwasser, Timm , journal =. 2023 , pages =

    work page 2023

  9. [9]

    Data-driven stochastic output-feedback predictive control:

    Pan, Guanru and Ou, Ruchuan and Faulwasser, Timm , booktitle =. Data-driven stochastic output-feedback predictive control:. 2023 , editor =

    work page 2023

  10. [10]

    Towards data-driven predictive control using wavelets , year =

    Sathyanarayanan, Kiran Kumar and Pan, Guanru and Faulwasser, Timm , journal =. Towards data-driven predictive control using wavelets , year =

    work page

  11. [11]

    Data-driven stochastic prediction with residual disturbances: application to residential building data , year =

    Guanru Pan and Dirk Peter Reinhardt and Sebastien Gros and Timm Faulwasser , note =. Data-driven stochastic prediction with residual disturbances: application to residential building data , year =

    work page

  12. [12]

    Towards fast computation of data-driven stochastic optimal control , year =

    Guanru Pan and Timm Faulwasser , note =. Towards fast computation of data-driven stochastic optimal control , year =

    work page

  13. [13]

    Towards a data-driven representation for nonlinear stochastic systems , year =

    Guanru Pan and Timm Faulwasser , note =. Towards a data-driven representation for nonlinear stochastic systems , year =

    work page

  14. [14]

    Batu and Bilgic, Deborah and Pan, Guanru and Koch, Alexander and Faulwasser, Timm , journal =

    Özmeteler, M. Batu and Bilgic, Deborah and Pan, Guanru and Koch, Alexander and Faulwasser, Timm , journal =. Data-driven uncertainty propagation for stochastic predictive control of multi-energy systems , year =

    work page

  15. [15]

    Data-Driven Control of Stochastic Systems: Representation, Prediction, and Optimal Control , year =

    Guanru Pan , school =. Data-Driven Control of Stochastic Systems: Representation, Prediction, and Optimal Control , year =

    work page

  16. [16]

    2023 , pages =

    Wang, Jiawei and Zheng, Yang and Li, Keqiang and Xu, Qing , journal =. 2023 , pages =

    work page 2023

  17. [17]

    Sullivan, T. J. , publisher =. 2015 , volume =

    work page 2015

  18. [18]

    and Shortt, Rae Michael , journal =

    Givens, Clark R. and Shortt, Rae Michael , journal =. A class of. 1984 , number =

    work page 1984

  19. [19]

    Willems, J. C. and Rapisarda, P. and Markovsky, I. and De Moor, B. L. M. , journal =. A note on persistency of excitation , year =

    work page

  20. [20]

    and Tesi, P

    De Persis, C. and Tesi, P. , journal =. Formulas for data-driven control: Stabilization, optimality, and robustness , year =

    work page

  21. [21]

    Stochastic model predictive control: An overview and perspectives for future research , year =

    Mesbah, Ali , journal =. Stochastic model predictive control: An overview and perspectives for future research , year =

    work page

  22. [22]

    Distributionally robust convex optimization , year =

    Wiesemann, Wolfram and Kuhn, Daniel and Sim, Melvyn , journal =. Distributionally robust convex optimization , year =

    work page

  23. [23]

    Bridging

    Nguyen, Viet Anh and Shafieezadeh-Abadeh, Soroosh and Kuhn, Daniel and Mohajerin Esfahani, Peyman , journal =. Bridging. 2023 , number =

    work page 2023

  24. [24]

    A distributionally robust optimization based method for stochastic model predictive control , year =

    Li, Bin and Tan, Yuan and Wu, Ai-Guo and Duan, Guang-Ren , journal =. A distributionally robust optimization based method for stochastic model predictive control , year =

    work page

  25. [25]

    , journal =

    Wiener, N. , journal =. The homogeneous chaos , year =

    work page

  26. [26]

    Field, R. V. and Grigoriu, M. , journal =. On the accuracy of the polynomial chaos approximation , year =

    work page

  27. [27]

    IEEE Control Syst

    M. IEEE Control Syst. Lett. , title =. 2017 , number =

    work page 2017

  28. [28]

    and Swarttouw, R

    Koekoek, R. and Swarttouw, R. F. , institution =. The. 1998 , address =

    work page 1998

  29. [29]

    and Karniadakis, G

    Xiu, D. and Karniadakis, G. E. , journal =. The. 2002 , number =

    work page 2002

  30. [30]

    Witteveen, J. A. S. and Bijl, H. , booktitle =. Modeling arbitrary uncertainties using. 2006 , address =

    work page 2006

  31. [31]

    , journal =

    Lefebvre, T. , journal =. On moment estimation from polynomial chaos expansion models , year =

    work page

  32. [32]

    Robust stability analysis of a simple data-driven model predictive control approach , year =

    Bongard, Joscha and Berberich, Julian and Köhler, Johannes and Allgöwer, Frank , journal =. Robust stability analysis of a simple data-driven model predictive control approach , year =

    work page

  33. [33]

    IFAC-Pap

    Berberich, Julian and K. IFAC-Pap. , title =. 2021 , number =

    work page 2021

  34. [34]

    On equivalence of data informativity for identification and data-driven control of partially observable systems , year =

    Sadamoto, Tomonori , journal =. On equivalence of data informativity for identification and data-driven control of partially observable systems , year =

    work page

  35. [35]

    On the certainty-equivalence approach to direct data-driven

    Dörfler, Florian and Tesi, Pietro and De Persis, Claudio , journal =. On the certainty-equivalence approach to direct data-driven. 2023 , number =

    work page 2023

  36. [36]

    and Giulioni, L

    Farina, M. and Giulioni, L. and Magni, L. and Scattolini, R. , booktitle =. A probabilistic approach to model predictive control , year =

    work page

  37. [37]

    Fristedt, B. E. and Gray, L. F. , publisher =. 2013 , file =

    work page 2013

  38. [38]

    and Hanebeck, Uwe D

    Weissel, Florian and Huber, Marco F. and Hanebeck, Uwe D. , booktitle =. Stochastic nonlinear model predictive control based on. 2009 , pages =

    work page 2009

  39. [39]

    The scenario approach for stochastic model predictive control with bounds on closed-loop constraint violations , year =

    Schildbach, Georg and Fagiano, Lorenzo and Frei, Christoph and Morari, Manfred , journal =. The scenario approach for stochastic model predictive control with bounds on closed-loop constraint violations , year =

    work page

  40. [40]

    Behavioral theory for stochastic systems?

    Faulwasser, Timm and Ou, Ruchuan and Pan, Guanru and Schmitz, Philipp and Worthmann, Karl , journal =. Behavioral theory for stochastic systems?. 2023 , pages =

    work page 2023

  41. [41]

    Calafiore, G. C. and Ghaoui, L. El , journal =. On distributionally robust chance-constrained linear programs , year =

    work page

  42. [42]

    and Li, S

    Yang, H. and Li, S. , booktitle =. A data-driven predictive controller design based on reduced hankel matrix , year =

    work page

  43. [43]

    Linear Algebra and Its Applications , title =

    Lobo, Miguel Sousa and Vandenberghe, Lieven and Boyd, Stephen and Lebret, Herv. Linear Algebra and Its Applications , title =. 1998 , number =

    work page 1998

  44. [44]

    2019 18th European Control Conference (ECC) , title =

    Coulson, Jeremy and Lygeros, John and D. 2019 18th European Control Conference (ECC) , title =. 2019 , organization =

    work page 2019

  45. [45]

    IEEE Trans

    Berberich, Julian and K. IEEE Trans. Autom. Control , title =. 2020 , number =

    work page 2020

  46. [46]

    and Plitt, K

    Bock, H. and Plitt, K. , journal =. A multiple shooting algorithm for direct solution of optimal control problems , year =

    work page

  47. [47]

    2022 , groups =

    Wu, Liang , journal =. 2022 , groups =

    work page 2022

  48. [48]

    Goulart and Eric C

    Paul J. Goulart and Eric C. Kerrigan and Jan M. Maciejowski , journal =. Optimization over state feedback policies for robust control with constraints , year =

    work page

  49. [49]

    Data-driven output prediction and control of stochastic systems: An innovation-based approach , year =

    Wang, Yibo and You, Keyou and Huang, Dexian and Shang, Chao , journal =. Data-driven output prediction and control of stochastic systems: An innovation-based approach , year =

    work page

  50. [50]

    and Falugi, Paola and Zagorowska, Marta and Shah, Nilay , journal =

    O’Dwyer, Edward and Kerrigan, Eric C. and Falugi, Paola and Zagorowska, Marta and Shah, Nilay , journal =. Data-driven predictive control with improved performance using segmented trajectories , year =

    work page

  51. [51]

    Data-driven distributionally robust bounds for stochastic model predictive control , year =

    Fochesato, Marta and Lygeros, John , booktitle =. Data-driven distributionally robust bounds for stochastic model predictive control , year =

    work page

  52. [52]

    and You, Fengqi , journal =

    Lu, Shuwen and Lee, Jay H. and You, Fengqi , journal =. Soft-constrained model predictive control based on data-driven distributionally robust optimization , year =

    work page

  53. [53]

    Data-driven distributionally robust

    Coppens, Peter and Patrinos, Panagiotis , journal =. Data-driven distributionally robust. 2021 , pages =

    work page 2021

  54. [54]

    2015 , number =

    IEEE Transactions on Automatic Control , title =. 2015 , number =

    work page 2015

  55. [55]

    Distributionally robust chance constrained data-enabled predictive control , year =

    Coulson, Jeremy and Lygeros, John and Dörfler, Florian , journal =. Distributionally robust chance constrained data-enabled predictive control , year =

    work page

  56. [56]

    and Smith, Stephen L

    Li, Ruiqi and Simpson-Porco, John W. and Smith, Stephen L. , journal =. Stochastic data-driven predictive control with equivalence to stochastic. 2023 , file =

    work page 2023

  57. [57]

    Markovsky, I. and D. Annu. Rev. Control , title =. 2021 , issn =

    work page 2021

  58. [58]

    Robust optimization , year =

    Ben-Tal, Aharon and El Ghaoui, Laurent and Nemirovski, Arkadi , publisher =. Robust optimization , year =

    work page

  59. [59]

    Real-time feasibility of data-driven predictive control for synchronous motor drives , year =

    Carlet, Paolo Gherardo and Favato, Andrea and Torchio, Riccardo and Toso, Francesco and Bolognani, Saverio and Dörfler, Florian , journal =. Real-time feasibility of data-driven predictive control for synchronous motor drives , year =

    work page

  60. [60]

    , journal =

    Willems, Jan C. , journal =. From time series to linear system—. 1986 , number =

    work page 1986

  61. [61]

    , journal =

    Willems, Jan C. , journal =. Paradigms and puzzles in the theory of dynamical systems , year =

    work page

  62. [62]

    2017 , number =

    Baggio, Giacomo and Sepulchre, Rodolphe , journal =. 2017 , number =

    work page 2017

  63. [63]

    , journal =

    Willems, Jan C. , journal =. Open stochastic systems , year =

    work page

  64. [64]

    IEEE Trans. Autom. Control , title =. 2024 , number =

    work page 2024

  65. [65]

    and Spanos, Pol D

    Ghanem, Roger G. and Spanos, Pol D. , publisher =. 2003 , edition =

    work page 2003

  66. [66]

    Uncertainty quantification via polynomial chaos expansion -- methods and applications for optimization of power systems , year =

    M. Uncertainty quantification via polynomial chaos expansion -- methods and applications for optimization of power systems , year =

    work page

  67. [67]

    Functional analysis,

    Br. Functional analysis,. 2011 , number =

    work page 2011

  68. [68]

    and Buehler, Edward A

    Paulson, Joel A. and Buehler, Edward A. and Mesbah, Ali , journal =. Arbitrary polynomial chaos for uncertainty propagation of correlated random variables in dynamic systems , year =

    work page

  69. [69]

    Data-driven uncertainty quantification using the arbitrary polynomial chaos expansion , year =

    Oladyshkin, Sergey and Nowak, Wolfgang , journal =. Data-driven uncertainty quantification using the arbitrary polynomial chaos expansion , year =

    work page

  70. [70]

    Sontag, E. D. , publisher =. Mathematical control theory: deterministic finite dimensional systems , year =

    work page

  71. [71]

    , journal =

    Wan, Yiming and Shen, Dongying Erin and Lucia, Sergio and Findeisen, Rolf and Braatz, Richard D. , journal =. Polynomial chaos-based. 2021 , pages =

    work page 2021

  72. [72]

    , journal =

    Wan, Yiming and Shen, Dongying and Lucia, Sergio and Findeisen, Rolf and Braatz, Richard D. , journal =. A polynomial chaos approach to robust. 2023 , number =

    work page 2023

  73. [73]

    On stochastic

    Fisher, James and Bhattacharya, Raktim , booktitle =. On stochastic. 2008 , organization =

    work page 2008

  74. [74]

    Machine Learning , title =

    H. Machine Learning , title =. 2021 , number =

    work page 2021

  75. [75]

    Real-time uncertainty decomposition for online learning control , year =

    Umlauft, Jonas and Lederer, Armin and Beckers, Thomas and Hirche, Sandra , note =. Real-time uncertainty decomposition for online learning control , year =

    work page

  76. [76]

    Noel A. C. Cressie , publisher =. 1996 , series =

    work page 1996

  77. [77]

    Mercer’s theorem on general domains: on the interaction between measures, kernels, and

    Steinwart, Ingo and Scovel, Clint , journal =. Mercer’s theorem on general domains: on the interaction between measures, kernels, and. 2012 , pages =

    work page 2012

  78. [78]

    , journal =

    Penrose, R. , journal =. A generalized inverse for matrices , year =

    work page

  79. [79]

    SIAM Journal on Matrix Analysis and Applications , title =

    El Ghaoui, Laurent and Lebret, Herv. SIAM Journal on Matrix Analysis and Applications , title =. 1997 , number =

    work page 1997

  80. [80]

    IEEE Trans

    Huang, Linbin and Zhen, Jianzhe and Lygeros, John and D. IEEE Trans. Autom. Control , title =. 2023 , number =

    work page 2023

Showing first 80 references.