Pith. sign in

REVIEW 1 major objections 5 minor 32 references

Compatibility of Multiple Control Barrier Functions for Constrained Nonlinear Systems

T0 review · 1 major / 5 minor · reviewed 2026-08-05 · deepseek-v4-flash

Pith's one-line read This paper proves that multiple exponential control barrier function constraints encoding box constraints on vector-valued outputs are always compatible whenever the output has a vector relative degree, and that the resulting quadratic-prog

desk verdict Solid extension of LTI multiple-CBF compatibility to nonlinear square systems; the compatibility and closed-form results hold up, but Theorem 2's forward-invariance claim needs a global existence assumption. read the letter →

arxiv 2509.04220 v1 pith:TD6OI2XJ submitted 2025-09-04 eess.SY cs.ROcs.SYmath.OC

classification eess.SYcs.ROcs.SYmath.OC
keywords controlbarrierfunctionsmultipleconstraintscompatibilityexponentialquadraticprogrammingsafetyfiltervectorrelativedegreeforwardinvariance
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 addresses a known gap in control barrier function (CBF) theory: most results handle one scalar state constraint, while real systems face several at once. The authors consider box constraints on the vector-valued output of a control-affine nonlinear system and encode each upper and lower bound as an exponential CBF (ECBF). Their central claim is that if the output has a vector relative degree, the 2m ECBF constraints are mutually compatible—there is always some input satisfying all of them—and the resulting quadratic-program safety filter is well-posed, locally Lipschitz, and available in closed form when the cost weight is chosen as the Gram matrix of the decoupling matrix. They also prove that this filter renders the safe set forward invariant and quantify how the filter degrades nominal tracking objectives. The practical payoff is a provably safe, optimization-free controller for simultaneous output constraints, demonstrated on a planar quadrotor.

What carries the argument

The vector relative degree of the output, captured by the m x m decoupling matrix B(x) whose rows are b_i(x)^T = L_g L_f^{r_i-1} y_i(x). Invertibility of B(x) on a neighborhood E of the safe set guarantees the rows b_i are linearly independent, which makes the pairwise opposite constraint directions ±b_i the only possible linear dependencies. The Farkas-type compatibility lemma (Lemma 1, following the proof of [11, Thm. 2]) then reduces compatibility to checking that a certifying linear combination of the constraint offsets is nonnegative; because each pair shares the same a_i and b_i with opposite signs, the cancellation leaves the positive gap y_i - y_i. The same linear independence also b

What would settle it

Find a system satisfying all other assumptions but where the decoupling matrix B(x) loses rank at a point inside the claimed safe set—for instance, the planar drone with θ reaching ±π/2—and exhibit a state in that set at which the two ECBF constraints on z_min and z_max admit no common input, or where the closed-form solution (27) violates one of the ECBF inequalities.

Watch

Extended reading notes

Core claim

For a square control-affine nonlinear system whose vector-valued output y has vector relative degree r = (r1,...,rm) on a neighborhood E of the safe set S, the 2m exponential CBF constraints encoding the box constraints y_i <= y_i(x) <= y_i are always compatible on E. Consequently, the QP safety filter (18) always has a feasible set, its solution u*(x) exists, is unique, and is locally Lipschitz continuous on E, and the closed-loop system keeps S forward invariant. When the quadratic cost weight is chosen as G(x) = B(x)^T B(x), where B(x) is the decoupling matrix, the filter admits the explicit closed-form solution (27): u*(x) = kd(x) + sum of active Lagrange-multiplier corrections, with eac

Load-bearing premise

The decoupling matrix B(x) must be invertible at every point in a neighborhood of the safe set; if the output loses its vector relative degree anywhere on the safe set, the rows b_i can become dependent, the compatibility proof fails, and the multiple box constraints can become mutually infeasible.

Editorial extensions

If this is right

  • Safety filters for multiple simultaneous output constraints no longer need online feasibility checks or conservative merging of constraints into a single CBF; compatibility comes from the system's relative-degree structure.
  • The closed-form filter (27) can run on platforms where an optimizer is impractical or prohibited, such as resource-constrained or aerospace systems.
  • For any system satisfying the vector-relative-degree assumption, the 2m ECBF constraints are always mutually feasible, so the QP is never infeasible and the controller is well-defined on the safe set.
  • The tracking analysis shows that the safety filter's effect on nominal tracking is input-to-state stable: tracking error degrades gracefully, bounded by a function of the filter's corrective action.
  • The approach extends the LTI compatibility results of [15] to a broader class of square nonlinear systems, subsuming several earlier special cases such as unicycle and fully actuated mechanical systems.

Reading between the lines

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

  • The proof structure suggests a general design recipe: choose auxiliary outputs as needed to achieve a vector relative degree, then every two-sided constraint on those outputs is automatically compatible. The planar-drone example shows this can force extra outputs (orientation) that consume actuation, so engineers will need to trade off which outputs to constrain directly.
  • The same compatibility argument may extend beyond boxes to any set of constraints where each constraint appears in opposite-sign pairs sharing a common high-order Lie derivative, e.g., symmetric limits on linear combinations of outputs—though the paper only proves the box case.
  • The closed-form multiplier λ_i*(x) doubles as an interpretable indicator of which constraint is active; in a deployed filter, these multipliers could be logged to detect which safety limits bind at any moment.
  • Since the paper explicitly leaves input bounds out of scope, a natural next test is whether compatibility persists when actuator saturation is added; the Farkas argument would need to be rechecked against the bounded-control case, where infeasibility typically appears at corners of the safe set.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

1 major / 5 minor

Summary. This paper studies multiple exponential control barrier functions (ECBFs) for MIMO affine nonlinear systems with box constraints on a vector-valued output. The main results are: when the output has a vector relative degree, the 2m ECBF constraints encoding the box constraints are mutually compatible (Lemma 2); the resulting QP safety filter (18) is locally Lipschitz (Lemma 3); with the particular weight G(x)=B(x)^T B(x), the QP admits the closed-form solution (27) (Theorem 3); and the safety filter degrades nominal tracking in an ISS sense (Proposition 1). The results are illustrated with planar quadrotor simulations.

Significance. The compatibility result is genuinely useful: it gives a clean structural condition under which multiple output-box ECBF constraints do not conflict, and the closed-form QP solution is valuable for embedded safety filters where online optimization is impractical. The Farkas/KKT derivations are largely self-contained and the case analysis in Theorem 3 is correct. The simulations demonstrate the framework. The principal defect is a missing forward-completeness condition in the safety theorem, which must be fixed before the practical safety guarantee is fully supported.

major comments (1)
  1. [§III, Theorem 2 and Lemma 4] Theorem 2 states that the controller from (18) renders S in (17) forward invariant and satisfies (25) for all t≥0. The proof only verifies the ECBF inequalities pointwise and then invokes Theorem 1, which requires the closed-loop solution to exist. No forward-completeness argument is given when S is unbounded. This is not a technicality: take ˙z=u, ˙q=q^2+u with output y=z, box z∈[0,1], relative degree r=1 (B=1), α=1, G=1, and kd=0. All conditions of Lemma 2 hold, and (18) yields u*=0 for every (z,q) with z∈[0,1]. Starting at (0.5,1), the closed-loop is ˙z=0, ˙q=q^2, so q(t)=1/(1−t) escapes at t=1 while z remains in [0,1]. Thus S={z∈[0,1]}×R is not forward invariant for t≥0. The theorem should either be stated on the maximal interval of existence, assume S is compact (as in Proposition 1), or assume the closed-loop system is forward complete. Lemma 4 inherits the same gap because it reli
minor comments (5)
  1. [§II-D, Lemma 2 proof] The indexing is confusing: 'j∈{1,3,...,2m−1}' with λ_j=λ_j and λ_{j+1}=λ_j mixes paired indices. Use explicit ̲underline and overbar notation (e.g., ̲λ_i and ̄λ_i) for the two constraints per output.
  2. [§III, Lemma 3 proof] Weierstrass' Theorem is invoked for existence over a closed but non-compact feasible set; it is the radial unboundedness/coercivity of the objective that guarantees existence. Also, the citation to Hager's theorem should explicitly verify the strong second-order sufficient condition (which follows from G(x)>0) and the LICQ at the solution (which follows from the independent b_i).
  3. [§III, Theorem 3 proof] The assertion that feasibility implies Slater's condition is not generally true for affine inequalities. In this problem Slater does hold, because the invertibility of B and y_i>y_i give a strict feasible point, but this should be stated rather than inferred from nonempty feasibility.
  4. [§III, Theorem 3 proof] Typo: 'complimentary slackness' should be 'complementary slackness.' In addition, 'the left-hand-side of the first inequality in (30)' should refer to (33), which is the inequality expressed in terms of ω_i, rather than the original KKT primal feasibility condition.
  5. [§V-A] The setpoint notation 'x=(1,1,04)' and 'x=(−1,0.2,04)' is ambiguous; write the six-dimensional state vector explicitly, e.g., x=(1,1,0,0,0,0).

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the main derivation is self-contained and does not reduce to its assumptions by construction.

full rationale

The paper's central chain is not circular. Lemma 2 establishes compatibility of the 2m ECBF constraints by combining the classical Farkas alternative (Lemma 1) with the invertibility of the decoupling matrix B(x) and the strict slack yi_bar > yi_under; the proof explicitly computes that a zero nonnegative combination of the constraint gradients forces the multipliers to pair up, and the resulting scalar sum is a positive multiple of yi_bar - yi_under. This is a direct mathematical consequence of the relative-degree assumption, not a fit or a definitional equivalence. Lemma 3's Lipschitz argument invokes Weierstrass and Hager's theorem for active-set Lipschitz continuity, neither of which imports the paper's conclusions. Theorem 3 selects G = B^T B, which diagonalizes the constraint Gram matrix, and then derives the closed-form controller from the KKT conditions; the expression in (27) is the exact solution of the QP (18), not a renamed input. Theorem 2's use of the standard ECBF theorem [22] and the intersection property of forward invariant sets [26] is likewise standard and externally supportable. The self-citations in the paper are limited: Lemma 1 cites [11] for Farkas's lemma, which is classical, and Section V-B's reduced-order-model example cites the authors' prior work [29], [30], [32], but that example is illustrative and not load-bearing for the paper's theoretical claims. Any concern about the forward-invariance theorem's failure to prove forward completeness for unbounded safe sets is a correctness or assumption issue, not a circularity issue; it does not involve the paper defining a result in terms of its own conclusion or fitting a parameter and calling it a prediction. Therefore the appropriate circularity score is 0.

Assumptions & free parameters 2 free parameters · 9 assumptions · 0 invented entities

The paper's central claim rests on the standard CBF/ECBF background theorems, the relative degree assumption, and the G-restriction for the closed form. No new physical entities are postulated. The list above records the design choices (alpha_i, G) and the implicit forward-completeness assumption.

free parameters (2)
  • ECBF gain vectors alpha_i = alpha_i = (1,2) in the example
    Chosen by the designer so the polynomial (5) has negative real roots; required alpha_i^1 > 0. Results hold for any valid alpha_i; in the example alpha=(1,2) gives roots -1,-1.
  • Cost weight G(x) = G(x)=B(x)^T B(x) in Theorem 3 and the examples
    The closed-form solution (27) holds only for this specific choice. For general G the solution remains the QP (18). This is a design choice, not fitted.
assumptions (9)
  • standard math Farkas' lemma / linear alternative theorem used in Lemma 1
    Used to convert compatibility to condition (13); proof follows [11, Thm. 2].
  • standard math KKT optimality conditions (with Slater's condition) for convex QP
    Used in Theorem 3 to verify the closed form; cited [27].
  • standard math Hager's theorem on Lipschitz continuity of parametric programs
    Used in Lemma 3 to show u* is locally Lipschitz; cited [24].
  • standard math Exponential CBF forward invariance theorem
    Theorem 1 ([22]) is the basis for forwarding S_i invariance in Theorem 2.
  • domain assumption Vector relative degree of output y on E containing S
    Definition 3 and Lemma 2; needed for independence of b_i and invertibility of B.
  • domain assumption Smoothness of f, g, y and local Lipschitzness of kd and G
    Assumed throughout (Section II); needed for local Lipschitz of u* and solution uniqueness.
  • ad hoc to paper Cost weight G(x)=B(x)^T B(x)
    Introduced in Theorem 3 to decouple the constraints and obtain the closed form; not needed for compatibility.
  • domain assumption For Proposition 1: safe set S compact and yd bounded with bounded derivatives
    Needed for the ISS bound (37).
  • domain assumption Existence of closed-loop solutions on [0,∞)
    Implicit in Theorem 2's 'for all t>=0'; not proven when S is unbounded.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Compatibility of Multiple Control Barrier Functions for Constrained Nonlinear Systems." pith.science (2026). https://pith.science/paper/TD6OI2XJ

@misc{pith2026250904220,
  author       = {Pith},
  title        = {Pith review of: Compatibility of Multiple Control Barrier Functions for Constrained Nonlinear Systems},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/TD6OI2XJ}},
  note         = {Machine review of arXiv:2509.04220}
}
read the original abstract

Control barrier functions (CBFs) are a powerful tool for the constrained control of nonlinear systems; however, the majority of results in the literature focus on systems subject to a single CBF constraint, making it challenging to synthesize provably safe controllers that handle multiple state constraints. This paper presents a framework for constrained control of nonlinear systems subject to box constraints on the systems' vector-valued outputs using multiple CBFs. Our results illustrate that when the output has a vector relative degree, the CBF constraints encoding these box constraints are compatible, and the resulting optimization-based controller is locally Lipschitz continuous and admits a closed-form expression. Additional results are presented to characterize the degradation of nominal tracking objectives in the presence of safety constraints. Simulations of a planar quadrotor are presented to demonstrate the efficacy of the proposed framework.

Figures

Figures reproduced from arXiv: 2509.04220 by the authors.

Figure 1
Figure 1. Closed-loop trajectories of the drone example from Sec. V-A [PITH_FULL_IMAGE:figures/full_fig_p007_1.png] view at source ↗
Figure 2
Figure 2. Closed-loop trajectories of the drone example from Sec. V-B [PITH_FULL_IMAGE:figures/full_fig_p008_2.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

32 extracted references · 31 canonical work pages

  1. [1]

    Control barrier function based quadratic programs for safety critical systems,

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

  2. [2]

    Borelli, A

    F. Borelli, A. Bemporad, and M. Morari, Predictive control for linear and hybrid systems . Cambridge University Press, 2017

  3. [3]

    Reference and command governors for systems with constraints: A survey on theory and applications,

    E. Garone, S. D. Cairano, and I. Kolmanovsky, “Reference and command governors for systems with constraints: A survey on theory and applications,” Automatica, vol. 75, pp. 306–328, 2017

  4. [4]

    Hamilton-Jacobi reachability: A brief overview and recent advances,

    S. Bansal, M. Chen, S. Herbert, and C. J. Tomlin, “Hamilton-Jacobi reachability: A brief overview and recent advances,” in Proc. Conf. Decis. Control, pp. 2242–2253, 2017

  5. [5]

    Data-driven safety filters: Hamilton-Jacobi reachability, control barrier functions, and predictive methods for uncertain systems,

    K. P. Wabersich, A. J. Taylor, J. J. Choi, K. Sreenath, C. J. Tom- lin, A. D. Ames, and M. N. Zeilinger, “Data-driven safety filters: Hamilton-Jacobi reachability, control barrier functions, and predictive methods for uncertain systems,” IEEE Control Systems Magazine , vol. 43, no. 5, pp. 137–177, 2023

  6. [6]

    The safety filter: A unified view of safety-critical control in autonomous systems,

    K. C. Hsu, H. Hu, and J. F. Fisac, “The safety filter: A unified view of safety-critical control in autonomous systems,” Annual Review of Control, Robotics, and Autonomous Systems , vol. 7, 2023

  7. [7]

    Towards a framework for realizable safety critical control through active set invariance,

    T. Gurriet, A. Singletary, J. Reher, L. Ciarletta, E. Feron, and A. Ames, “Towards a framework for realizable safety critical control through active set invariance,” in Proc. ACM/IEEE Int. Conf. Cyber-Physical Syst., pp. 98–106, 2018

  8. [8]

    Safety barrier certificates for collisions-free multirobot systems,

    L. Wang, A. D. Ames, and M. Egerstedt, “Safety barrier certificates for collisions-free multirobot systems,” IEEE Trans. Robot , vol. 33, no. 3, pp. 661–674, 2017

Show all 32 references
  1. [9]

    Nonsmooth barrier func- tions with applications to multi-robot systems,

    P. Glotfelter, J. Cort ´es, and M. Egerstedt, “Nonsmooth barrier func- tions with applications to multi-robot systems,” IEEE Contr. Syst. Lett., vol. 1, no. 2, pp. 310–315, 2017

  2. [10]

    Control barrier functions for signal temporal logic tasks,

    L. Lindemann and D. V . Dimarogonas, “Control barrier functions for signal temporal logic tasks,” IEEE Contr. Syst. Lett. , vol. 3, no. 1, pp. 96–101, 2019

  3. [11]

    Composing control barrier functions for complex safety specifications,

    T. G. Molnar and A. D. Ames, “Composing control barrier functions for complex safety specifications,” IEEE Contr. Syst. Lett. , vol. 7, pp. 3615–3620, 2023

  4. [12]

    Adaptation for validation of consolidated control barrier functions,

    M. Black and D. Panagou, “Adaptation for validation of consolidated control barrier functions,” in Proc. Conf. Decis. Control, pp. 751–757, 2023

  5. [13]

    Constrained control of input–output linearizable systems using control sharing barrier functions,

    X. Xu, “Constrained control of input–output linearizable systems using control sharing barrier functions,” Automatica, vol. 87, 2018

  6. [14]

    A robust, multiple control barrier function framework for input constrained systems,

    W. S. Cortez, X. Tan, and D. V . Dimarogonas, “A robust, multiple control barrier function framework for input constrained systems,” IEEE Contr. Syst. Lett. , no. 6, pp. 1742–1747, 2022

  7. [15]

    Servo-controllers for linear time- invariant systems with operational constraints,

    E. Lavretsky and M. Menner, “Servo-controllers for linear time- invariant systems with operational constraints,” in Proc. Amer. Control Conf., 2025. to appear

  8. [16]

    Safe-by-design control for Euler–Lagrange systems,

    W. S. Cortez and D. V . Dimarogonas, “Safe-by-design control for Euler–Lagrange systems,” Automatica, vol. 146, p. 110620, 2022

  9. [17]

    Compatibility checking of multiple control barrier functions for input constrained systems,

    X. Tan and D. V . Dimarogonas, “Compatibility checking of multiple control barrier functions for input constrained systems,” in Proc. Conf. Decis. Control, pp. 939–944, 2022

  10. [18]

    Compositions of multiple control barrier functions under input constraints,

    J. Breeden and D. Panagou, “Compositions of multiple control barrier functions under input constraints,” in Proc. Amer. Control Conf. , pp. 3688–3695, 2023

  11. [19]

    On the feasibility and continuity of feedback controllers defined by multiple control barrier functions,

    A. Isaly, M. Ghanbarpour, R. G. Sanfelice, and W. E. Dixon, “On the feasibility and continuity of feedback controllers defined by multiple control barrier functions,” IEEE Trans. Autom. Control, vol. 69, no. 11, pp. 7326–7339, 2024

  12. [20]

    Isidori, Nonlinear Control Systems

    A. Isidori, Nonlinear Control Systems . Springer, third ed., 1995

  13. [21]

    Robust control barrier functions for constrained sta- bilization of nonlinear systems,

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

  14. [22]

    Exponential control barrier functions for enforcing high relative-degree safety-critical constraints,

    Q. Nguyen and K. Sreenath, “Exponential control barrier functions for enforcing high relative-degree safety-critical constraints,” in Proc. Amer. Control Conf., pp. 322–328, 2016

  15. [23]

    D. P. Bertsekas, Nonlinear Programming. Athena Scientific, third ed., 2016

  16. [24]

    Lipschitz continuity of constrained processes,

    W. W. Hager, “Lipschitz continuity of constrained processes,” SIAM J. Control and Optimization , vol. 17, no. 3, pp. 321–338, 1979

  17. [25]

    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 Contr. Syst. Lett. , vol. 7, pp. 127–132, 2023

  18. [26]

    Blanchini and S

    F. Blanchini and S. Miani, Set-theoretic methods in control . Springer, 2008

  19. [27]

    Boyd and L

    S. Boyd and L. Vandenberghe, Convex Optimization . Cambridge University Press, 2004

  20. [28]

    H. K. Khalil, Nonlinear Systems. Prentice Hall, 3 ed., 2002

  21. [29]

    Safety-critical control for autonomous systems: Control barrier functions via reduced order models,

    M. H. Cohen, T. G. Molnar, and A. D. Ames, “Safety-critical control for autonomous systems: Control barrier functions via reduced order models,” Annual Reviews in Control , vol. 57, p. 100947, 2024

  22. [30]

    Safety-critical controller synthesis with reduced-order models,

    M. H. Cohen, N. Csomay-Shanklin, W. D. Compton, T. G. Molnar, and A. D. Ames, “Safety-critical controller synthesis with reduced-order models,” in Proc. Amer. Control Conf., pp. 5216–5221, 2025

  23. [31]

    Minimum snap trajectory generation and control for quadrotors,

    D. Mellinger and V . Kumar, “Minimum snap trajectory generation and control for quadrotors,” in Proc. Int. Conf. Robot. and Autom. , pp. 2520–2525, 2011

  24. [32]

    Control barrier function synthesis for nonlinear systems with dual relative degree,

    G. Bahati, R. K. Cosner, M. H. Cohen, R. M. Bena, and A. D. Ames, “Control barrier function synthesis for nonlinear systems with dual relative degree,” arXiv preprint arXiv:2504.00397 , 2025. ACKNOWLEDGMENTS We gratefully acknowledge Dr. Heather Hussain, whose continuous feedb...

Pith tools

Reviewed August 5, 2026 · model on record in the stance chip above.