pith. the verified trust layer for science. sign in

arxiv: 2604.19893 · v1 · submitted 2026-04-21 · 📡 eess.SY · cs.SY

Output Feedback Backup Control Barrier Functions: Safety Guarantees Under Input Bounds and State Estimation Error

David E. J. van Wijk , Tamas G. Molnar , Samuel Coogan , Manoranjan Majji , Aaron D. Ames , Joel W. Burdick This is my paper

Pith reviewed 2026-05-10 01:25 UTC · model grok-4.3

classification 📡 eess.SY cs.SY
keywords control barrier functionsbackup control barrier functionsoutput feedbackstate estimation errorinput constraintssafety-critical controluncertainty envelopes
0
0 comments X p. Extension

The pith

Using an uncertainty envelope around the estimated flow lets output-feedback backup barriers guarantee that the true state stays safe despite estimation error and input bounds.

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

The paper develops output feedback backup control barrier functions that handle the fact that only an estimated state is known and that control is applied to that estimate. It constructs an uncertainty envelope centered on the flow computed from the estimate and shows that keeping the envelope inside the safe set forces the unknown true state to remain safe as well. The construction accounts for the coupling created when the true dynamics evolve under feedback from the estimate. It further proves that a control input satisfying the input bounds always exists to enforce this envelope safety. Readers care because real controllers rarely have perfect state information yet must still avoid unsafe regions.

Core claim

We propose Output Feedback Backup Control Barrier Functions (O-bCBFs) defined on an uncertainty envelope centered around the estimated flow. Ensuring safety of this envelope guarantees that the true state satisfies the safety constraints. In the presence of state uncertainty, the resulting O-bCBFs admit a feasible control input that guarantees safety of the true state even under input constraints.

What carries the argument

The uncertainty envelope centered on the estimated flow, which over-approximates every possible true trajectory arising from feedback applied to the estimate.

If this is right

  • Safety of the true state follows directly once the envelope is kept safe.
  • A feasible bounded input always exists that renders the envelope invariant.
  • The approach handles the coupling between estimated-state feedback and unknown true dynamics without requiring the true state.
  • The same envelope construction works for any controller that depends only on the estimate.

Where Pith is reading between the lines

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

  • The same envelope idea could be paired with any observer that supplies explicit error bounds rather than requiring a specific estimator.
  • Extensions to systems whose dynamics are only approximately known would require inflating the envelope by an additional robustness margin.
  • The existence result suggests that O-bCBFs can be combined with quadratic programs for real-time implementation without losing feasibility guarantees.
  • Multi-agent settings where each agent sees only a local estimate become feasible once each agent maintains its own envelope.

Load-bearing premise

The uncertainty envelope correctly contains every possible true flow that results when control is computed from the estimated state.

What would settle it

A concrete trajectory in which the true state leaves the safe set while the uncertainty envelope remains inside it, or a case in which no bounded input keeps the envelope safe when the true system can still be made safe.

Figures

Figures reproduced from arXiv: 2604.19893 by Aaron D. Ames, David E. J. van Wijk, Joel W. Burdick, Manoranjan Majji, Samuel Coogan, Tamas G. Molnar.

Figure 1
Figure 1. Figure 1: Sketch of open-loop safety method for the presented output [PITH_FULL_IMAGE:figures/full_fig_p006_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Illustration of closed-loop safety method for the presented [PITH_FULL_IMAGE:figures/full_fig_p010_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Double integrator system (52) under safe output feedback, [PITH_FULL_IMAGE:figures/full_fig_p012_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: The safe control input (green) satisfies input constraints, [PITH_FULL_IMAGE:figures/full_fig_p012_4.png] view at source ↗
Figure 6
Figure 6. Figure 6: Desired control (dashed) and safe control signal from [PITH_FULL_IMAGE:figures/full_fig_p013_6.png] view at source ↗
read the original abstract

Guaranteeing the safety of controllers is vital for real-world applications, but is markedly difficult when the states are not perfectly known and when the control inputs are bounded. Backup control barrier functions (bCBFs) use predictions of the flow under a prescribed controller to achieve safety in the presence of bounded inputs and perfect state information. However, when only an estimate of the true state is known, this flow may not be precisely computed, as the initial condition is unknown. Furthermore, the true flow evolves using feedback from the estimated state, thus introducing coupling between known and unknown flows. To address these challenges, we propose a technique that leverages an uncertainty envelope centered around the estimated flow and show that ensuring the safety of this envelope guarantees that the true state satisfies the safety constraints. Additionally, we show that in the presence of state uncertainty, using the resulting Output Feedback Backup Control Barrier Functions (O-bCBFs), there always exists a feasible control input that can guarantee the safety of the true state, even in the presence of input constraints.

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 / 1 minor

Summary. The paper extends backup control barrier functions (bCBFs) to output-feedback settings with state estimation error and input bounds. It constructs an uncertainty envelope around the estimated flow (under feedback from the estimated state) and defines Output Feedback bCBFs (O-bCBFs) such that safety of the envelope guarantees safety of the true state; it further claims that a feasible control input always exists to enforce this safety despite input constraints.

Significance. If the envelope construction and existence result are rigorous, the work fills a practical gap in safe control under imperfect state information and actuator saturation, directly building on prior bCBF literature. The existence of feasible controls is a notable strength, as it indicates the method remains non-vacuous under constraints.

major comments (2)
  1. Abstract and uncertainty-envelope construction: The central claim requires that the envelope (centered on the estimated flow under estimated-state feedback) contains every true trajectory consistent with bounded initial error and the same bounded control sequence. Because the control law is evaluated at the estimate, input saturation can differ along true vs. estimated trajectories; the manuscript must explicitly derive or bound the worst-case saturation effect and the Lipschitz growth of the error under bounded inputs, or the inclusion fails even with known estimation-error bounds.
  2. Safety-implication theorem (likely the main result following the envelope definition): The assertion that envelope safety implies true-state safety is load-bearing and must be proved with explicit assumptions on the dynamics (Lipschitz constants, input bounds) and a propagation argument for the envelope; the abstract supplies no sketch, leaving open whether the coupling between estimated and true flows is fully accounted for.
minor comments (1)
  1. Notation: Distinguish the estimated flow, true flow, and envelope bounds more clearly in definitions and any accompanying figures to prevent reader confusion about which trajectory is being bounded.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive comments that highlight important aspects of rigor in the envelope construction and safety proof. We address each point below and have revised the manuscript to make the derivations and assumptions fully explicit.

read point-by-point responses

  1. Referee: Abstract and uncertainty-envelope construction: The central claim requires that the envelope (centered on the estimated flow under estimated-state feedback) contains every true trajectory consistent with bounded initial error and the same bounded control sequence. Because the control law is evaluated at the estimate, input saturation can differ along true vs. estimated trajectories; the manuscript must explicitly derive or bound the worst-case saturation effect and the Lipschitz growth of the error under bounded inputs, or the inclusion fails even with known estimation-error bounds.

    Authors: We agree that explicit bounds on saturation mismatch and error growth are necessary for the envelope inclusion to hold. In the full manuscript (Section III-B), the envelope is constructed by propagating the maximum possible deviation using the system Lipschitz constant L and the input bound u_max; the worst-case saturation difference is bounded by considering the maximum state error at each instant and the Lipschitz continuity of the saturation function. We have added a supporting lemma (Lemma 2) that states these bounds explicitly and proves the envelope contains all admissible true trajectories. The revised abstract now includes a one-sentence sketch of this construction. revision: yes

  2. Referee: Safety-implication theorem (likely the main result following the envelope definition): The assertion that envelope safety implies true-state safety is load-bearing and must be proved with explicit assumptions on the dynamics (Lipschitz constants, input bounds) and a propagation argument for the envelope; the abstract supplies no sketch, leaving open whether the coupling between estimated and true flows is fully accounted for.

    Authors: The safety implication is proved in Theorem 1. The proof explicitly assumes f is L-Lipschitz, inputs are bounded by u_max, and initial estimation error is bounded by δ; it employs a Gronwall-type propagation argument over the finite backup horizon to show that any true trajectory starting inside the envelope and driven by the same (possibly saturated) input sequence remains inside the envelope. The coupling induced by feedback from the estimate is accounted for by using the identical control sequence in both the estimated and true flows when deriving the envelope radius. We have expanded the theorem statement to list all assumptions, added the propagation steps in full detail, and inserted a brief proof sketch into the abstract. revision: yes

Circularity Check

0 steps flagged

No circularity: derivation relies on independent set-theoretic arguments

full rationale

The paper extends backup CBFs to output-feedback settings by defining an uncertainty envelope around the estimated flow under bounded inputs and known estimation-error bounds. No equations in the provided abstract or described chain reduce a claimed prediction or existence result to a fitted parameter or self-referential definition. Prior bCBF results are cited for the perfect-information case, but the central guarantee (safety of the envelope implies safety of the true state, plus existence of a feasible input) is derived from the envelope construction itself rather than from a self-citation chain or uniqueness theorem imported from the same authors. The derivation therefore remains self-contained against external benchmarks and does not collapse to its inputs by construction.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 1 invented entities

Abstract-only review prevents exhaustive audit; the approach implicitly relies on standard control-theoretic assumptions and introduces the uncertainty envelope as a new bounding construct.

axioms (2)
  • domain assumption System dynamics are known and sufficiently regular (e.g., Lipschitz) to compute flows under a prescribed controller
    Required to define both the estimated flow and the uncertainty envelope.
  • domain assumption Bounds on state estimation error are known a priori
    Necessary to construct a valid uncertainty envelope that contains the true trajectory.
invented entities (1)
  • Uncertainty envelope no independent evidence
    purpose: Set of possible true trajectories consistent with the estimated state and error bounds
    New bounding object introduced to decouple known and unknown flows under output feedback.

pith-pipeline@v0.9.0 · 5509 in / 1229 out tokens · 33113 ms · 2026-05-10T01:25:10.223190+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 · 4 canonical work pages

  1. [1]

    Control barrier function based quadratic programs for safety critical sys- tems,

    A. D. Ames, X. Xu, J. W. Grizzle, and P. Tabuada, “Control barrier function based quadratic programs for safety critical sys- tems,” IEEE Trans. Autom. Control , vol. 62, no. 8, pp. 3861– 3876, 2017

    2017

  2. [2]

    Control barrier functions for complete and incom- plete information stochastic systems,

    A. Clark, “Control barrier functions for complete and incom- plete information stochastic systems,” in 2019 Proc. Amer. Control Conf. (ACC) , pp. 2928–2935, 2019

    2019

  3. [3]

    Risk-bounded control with Kalman filtering and stochastic barrier functions,

    S. Yaghoubi, G. Fainekos, T. Yamaguchi, D. Prokhorov, and B. Hoxha, “Risk-bounded control with Kalman filtering and stochastic barrier functions,” in 2021 60th IEEE Conf. Decision and Control (CDC) , pp. 5213–5219, 2021

    2021

  4. [4]

    Risk-bounded control using 13 stochastic barrier functions,

    S. Yaghoubi, K. Majd, G. Fainekos, T. Yamaguchi, D. Prokhorov, and B. Hoxha, “Risk-bounded control using 13 stochastic barrier functions,” IEEE Control Syst. Lett. , vol. 5, no. 5, pp. 1831–1836, 2021

    2021

  5. [5]

    Belief control barrier func- tions for risk-aware control,

    M. Vahs, C. Pek, and J. Tumova, “Belief control barrier func- tions for risk-aware control,” IEEE Robot. Aut. Lett. , vol. 8, no. 12, pp. 8565–8572, 2023

    2023

  6. [6]

    Risk-aware control for robots with non-Gaussian belief spaces,

    M. Vahs and J. Tumova, “Risk-aware control for robots with non-Gaussian belief spaces,” in 2024 Proc. Int. Conf. Robot. and Autom. (ICRA) , pp. 11661–11667, IEEE, 2024

    2024

  7. [7]

    Safe perception-based control under stochastic sensor uncertainty us- ing conformal prediction,

    S. Yang, G. J. Pappas, R. Mangharam, and L. Lindemann, “Safe perception-based control under stochastic sensor uncertainty us- ing conformal prediction,” in 2023 62nd IEEE Conf. Decision and Control (CDC) , pp. 6072–6078, 2023

    2023

  8. [8]

    Robust control barrier functions for constrained stabilization of nonlinear systems,

    M. Jankovic, “Robust control barrier functions for constrained stabilization of nonlinear systems,” Automatica, vol. 96, pp. 359–367, Oct. 2018

    2018

  9. [9]

    Robust Control Barrier Functions under high relative degree and input constraints for satellite trajectories,

    J. Breeden and D. Panagou, “Robust Control Barrier Functions under high relative degree and input constraints for satellite trajectories,” Automatica, vol. 155, p. 111109, Sept. 2023

    2023

  10. [10]

    Parame- terized barrier functions to guarantee safety under uncertainty,

    A. Alan, T. G. Molnar, A. D. Ames, and G. Orosz, “Parame- terized barrier functions to guarantee safety under uncertainty,” IEEE Control Systems Letters , vol. 7, pp. 2077–2082, 2023

    2077

  11. [11]

    Robust control barrier functions using uncertainty estimation with application to mobile robots,

    E. Da¸ s and J. W. Burdick, “Robust control barrier functions using uncertainty estimation with application to mobile robots,” IEEE Trans. Autom. Control , pp. 1–8, 2025

    2025

  12. [12]

    Disturbance observer-based robust con- trol barrier functions,

    Y. Wang and X. Xu, “Disturbance observer-based robust con- trol barrier functions,” in Proc. Amer. Control Conf. , pp. 3681– 3687, 2023

    2023

  13. [13]

    Robust adap- tive control barrier functions: An adaptive and data-driven approach to safety,

    B. T. Lopez, J.-J. E. Slotine, and J. P. How, “Robust adap- tive control barrier functions: An adaptive and data-driven approach to safety,” IEEE Control Syst. Lett. , vol. 5, no. 3, pp. 1031–1036, 2021

    2021

  14. [14]

    Robust adaptive control barrier functions for input-affine systems: Ap- plication to uncertain manipulator safety constraints,

    D. Zeng, Y. Jiang, Y. Wang, H. Zhang, and Y. Feng, “Robust adaptive control barrier functions for input-affine systems: Ap- plication to uncertain manipulator safety constraints,” IEEE Control Syst. Lett. , vol. 8, pp. 279–284, 2024

    2024

  15. [15]

    Safe and Robust Observer- Controller Synthesis Using Control Barrier Functions,

    D. R. Agrawal and D. Panagou, “Safe and Robust Observer- Controller Synthesis Using Control Barrier Functions,” IEEE Control Syst. Lett. , vol. 7, pp. 127–132, 2023

    2023

  16. [16]

    Observer-based control barrier functions for safety critical systems,

    Y. Wang and X. Xu, “Observer-based control barrier functions for safety critical systems,” in 2022 Proc. Amer. Control Conf. (ACC), pp. 709–714, IEEE, 2022

    2022

  17. [17]

    Guaranteeing safety of learned perception modules via measurement-robust control barrier functions,

    S. Dean, A. J. Taylor, R. K. Cosner, B. Recht, and A. D. Ames, “Guaranteeing safety of learned perception modules via measurement-robust control barrier functions,” Conf. Robot Learning, 2020

    2020

  18. [18]

    Learning robust output control barrier functions from safe expert demonstrations,

    L. Lindemann, A. Robey, L. Jiang, S. Das, S. Tu, and N. Matni, “Learning robust output control barrier functions from safe expert demonstrations,” IEEE Open J. Control Syst. , vol. 3, pp. 158–172, 2024

    2024

  19. [19]

    Safe navigation under state un- certainty: Online adaptation for robust control barrier func- tions,

    E. Das, R. Nanayakkara, X. Tan, R. M. Bena, J. W. Burdick, P. Tabuada, and A. D. Ames, “Safe navigation under state un- certainty: Online adaptation for robust control barrier func- tions,” arXiv preprint arXiv:2508.19159 , 2025

    work page arXiv 2025

  20. [20]

    Safety un- der state uncertainty: Robustifying control barrier functions,

    R. Nanayakkara, A. D. Ames, and P. Tabuada, “Safety un- der state uncertainty: Robustifying control barrier functions,” arXiv preprint arXiv:2508.17226 , 2025

    work page arXiv 2025

  21. [21]

    An Online Approach to Active Set Invariance,

    T. Gurriet, M. Mote, A. D. Ames, and E. Feron, “An Online Approach to Active Set Invariance,” in 2018 IEEE Conf. Deci- sion and Control (CDC) , pp. 3592–3599, Dec. 2018

    2018

  22. [22]

    A scalable safety critical control framework for nonlinear systems,

    T. Gurriet, M. Mote, A. Singletary, P. Nilsson, E. Feron, and A. D. Ames, “A scalable safety critical control framework for nonlinear systems,” IEEE Access, vol. 8, pp. 187249–187275, 2020

    2020

  23. [23]

    Compar- ing run time assurance approaches for safe spacecraft docking,

    K. Dunlap, M. Hibbard, M. Mote, and K. Hobbs, “Compar- ing run time assurance approaches for safe spacecraft docking,” IEEE Control Syst. Lett. , vol. 6, pp. 1849–1854, 2022

    2022

  24. [24]

    Runtime assurance for safety-critical systems: An introduction to safety filtering approaches for complex control systems,

    K. L. Hobbs, M. L. Mote, M. C. Abate, S. D. Coogan, and E. M. Feron, “Runtime assurance for safety-critical systems: An introduction to safety filtering approaches for complex control systems,” IEEE Contr. Syst. Mag. , vol. 43, no. 2, pp. 28–65, 2023

    2023

  25. [25]

    A backup control barrier function approach for safety-critical control of mechanical systems under multiple constraints,

    D. Ko and W. K. Chung, “A backup control barrier function approach for safety-critical control of mechanical systems under multiple constraints,” IEEE/ASME Trans. Mechatron., pp. 1– 12, 2024

    2024

  26. [26]

    A learning-based framework for safe human- robot collaboration with multiple backup control barrier func- tions,

    N. C. Janwani, E. Da¸ s, T. Touma, S. X. Wei, T. G. Molnar, and J. W. Burdick, “A learning-based framework for safe human- robot collaboration with multiple backup control barrier func- tions,” in 2024 Proc. Int. Conf. Robot. and Autom. (ICRA) , pp. 11676–11682, IEEE, 2024

    2024

  27. [27]

    Soft-minimum and soft-maximum barrier functions for safety with actuation constraints,

    P. Rabiee and J. B. Hoagg, “Soft-minimum and soft-maximum barrier functions for safety with actuation constraints,” Auto- matica, vol. 171, p. 111921, 2025

    2025

  28. [28]

    Forward and control invariance analysis of backup control barrier function induced safe sets for online safety of nonlinear systems,

    P. M. Rivera, “Forward and control invariance analysis of backup control barrier function induced safe sets for online safety of nonlinear systems,” in 63rd Conf. Decision and Con- trol, pp. 8150–8157, 2024

    2024

  29. [29]

    Measurement-robust control barrier functions: Certainty in safety with uncertainty in state,

    R. K. Cosner, A. W. Singletary, A. J. Taylor, T. G. Mol- nar, K. L. Bouman, and A. D. Ames, “Measurement-robust control barrier functions: Certainty in safety with uncertainty in state,” in 2021 IEEE/RSJ Int. Conf. Intell. Robots Syst. (IROS), pp. 6286–6291, IEEE, 2021

    2021

  30. [30]

    Braking within barriers: Constructive safety-critical control for input- constrained vehicles via the backup set method,

    L. Gacsi, A. K. Kiss, and T. G. Molnar, “Braking within barriers: Constructive safety-critical control for input- constrained vehicles via the backup set method,” arXiv preprint arXiv:2510.15797, 2025

    work page arXiv 2025

  31. [31]

    Safety-critical control with bounded inputs via reduced order models,

    T. G. Molnar and A. D. Ames, “Safety-critical control with bounded inputs via reduced order models,” in 2023 Proc. Amer. Control Conf. (ACC) , pp. 1414–1421, 2023

    2023

  32. [32]

    Uncertainty estimators for robust backup control barrier functions,

    D. E. J. van Wijk, E. Da¸ s, A. Alan, S. Coogan, T. G. Mol- nar, J. W. Burdick, M. Majji, and K. L. Hobbs, “Uncertainty estimators for robust backup control barrier functions,” arXiv preprint arXiv:2503.15734, 2025

    work page arXiv 2025

  33. [33]

    Disturbance-robust backup control barrier functions: Safety under uncertain dynamics,

    D. E. J. van Wijk, S. Coogan, T. G. Molnar, M. Majji, and K. L. Hobbs, “Disturbance-robust backup control barrier functions: Safety under uncertain dynamics,” IEEE Control Syst. Lett. , vol. 8, pp. 2817–2822, 2024

    2024

  34. [34]

    Confidence-aware safe and stable control of control-affine systems,

    S. Wei, P. Krishnamurthy, and F. Khorrami, “Confidence-aware safe and stable control of control-affine systems,” in 2024 Proc. Amer. Control Conf. (ACC) , pp. 3371–3376, 2024

    2024

  35. [35]

    H. K. Khalil, Nonlinear control. Boston: Pearson, 1st ed., 2015

    2015

  36. [36]

    An EKF-based nonlinear observer with a prescribed degree of stability,

    K. Reif, F. Sonnemann, and R. Unbehauen, “An EKF-based nonlinear observer with a prescribed degree of stability,” Auto- matica, vol. 34, no. 9, pp. 1119–1123, 1998

    1998

  37. [37]

    Bullo, Contraction Theory for Dynamical Systems

    F. Bullo, Contraction Theory for Dynamical Systems . Kindle Direct Publishing, 1.2 ed., 2024

    2024

  38. [38]

    Sontag, Input to state stability: Basic concepts and results , pp

    E. Sontag, Input to state stability: Basic concepts and results , pp. 163–220. Lecture Notes in Mathematics, Germany: Springer Verlag, 2008

    2008

  39. [39]

    Nonlinear observers robust to mea- surement disturbances in an ISS sense,

    H. Shim and D. Liberzon, “Nonlinear observers robust to mea- surement disturbances in an ISS sense,” IEEE Trans. Autom. Control, vol. 61, no. 1, pp. 48–61, 2016

    2016

  40. [40]

    R. A. Freeman and P. Kokotovi´ c, Robust Nonlinear Control Design. Birkh¨ auser Boston, 1996

    1996

  41. [41]

    Khalil, Nonlinear Systems

    H. Khalil, Nonlinear Systems . Pearson Education, Prentice Hall, 2 ed., 2002

    2002

  42. [42]

    Input-to-state safety with con- trol barrier functions,

    S. Kolathaya and A. D. Ames, “Input-to-state safety with con- trol barrier functions,” IEEE Control Syst. Lett. , vol. 3, no. 1, pp. 108–113, 2019

    2019

  43. [43]

    Nocedal and S

    J. Nocedal and S. J. Wright, Numerical optimization. Springer series in operations research, New York: Springer, 2nd ed., 2006

    2006

  44. [44]

    E. D. Sontag, Contractive Systems with Inputs , pp. 217–228. Berlin, Heidelberg: Springer Berlin Heidelberg, 2010

    2010

  45. [45]

    Blanchini and S

    F. Blanchini and S. Miani, Set-Theoretic Methods in Con- trol. Systems & Control: Foundations & Applications, Cham: Springer International Publishing, 2015

    2015

  46. [46]

    Perspectives on contractivity in control, optimization, and learning,

    A. Davydov and F. Bullo, “Perspectives on contractivity in control, optimization, and learning,” IEEE Control Syst. Lett. , vol. 8, pp. 2087–2098, 2024

    2087

  47. [47]

    M. L. Mote, Optimization-based approaches to safety-critical control with applications to space systems . PhD thesis, Georgia Institute of Technology, 2021. 14

    2021