Navigating Polytopes with Safety: A Control Barrier Function Approach

Tamas G. Molnar

REVIEW 3 major objections 4 minor 28 references

Navigating Polytopes with Safety: A Control Barrier Function Approach

T0 review · 3 major / 4 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read This paper constructs a closed-form control barrier function whose superlevel set under-approximates the safe region, so enforcing the barrier condition provably prevents collisions between a polytope agent and a polytope environment.

desk verdict A genuinely useful closed-form polytope CBF construction, with a load-bearing but fixable gap: the b=0 buffer choice does not guarantee safety, and the CBF condition is never verified. read the letter →

arxiv 2505.17270 v1 pith:Y2P3ZRJE submitted 2025-05-22 eess.SY cs.ROcs.SY

Tamas G. Molnar This is my paper · ORCID

classification eess.SYcs.ROcs.SY

keywords controlbarrierfunctionscollision-freenavigationpolytopeagentenvironmentsigneddistanceunder-approximationlog-sum-expsmoothingsafety-criticalclosed-formcontrollers

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 aims to establish a systematic, closed-form construction of control barrier functions (CBFs) for an agent with polytope geometry moving safely through a polytope environment, with controllers that come with formal collision-free guarantees. The key idea is to express safety as a nested max-min of half-space constraints over the agent's vertices and the environment's convex pieces, then smooth that expression with log-sum-exp so it is differentiable and usable as a CBF. The paper proves that the unsmoothed vertex condition implies the agent's entire convex hull lies in the safe region, and argues that the smoothed version with a buffer under-approximates the safe set, so that enforcing the CBF condition yields provably collision-free motion. If successful, the approach would give real-time safety filters for robotics and autonomous vehicles that require only evaluating one closed-form function, not solving an optimization online.

What carries the argument

The central object is the vertex-based under-approximation $\varphi$(x) = max_{j in J} min_{i in I_j} min_{k in K} psi_i(p_k) in Eq. (31), where psi_i are the affine half-space barrier functions describing the environment, k indexes the agent's vertices, and j indexes the convex polytopes whose union is the environment. Theorem 2 shows phi>=0 implies the convex hull of the agent is contained in the environment, because each wall barrier is linear in the vertex positions and convex combinations preserve nonnegativity. The smooth companion h in Eq. (36) replaces the nested max and min by log-sum-exp sums, giving a continuously differentiable function whose superlevel set is intended to under-approximate the safe set, with buffer b absorbing the smoothing error; enforcing the CBF condition on h is what generates the safe controller.

What would settle it

Evaluate h in Eq. (36) with b=0 for a point or polytope agent in a non-convex polytope where two environment pieces meet at a concave corner, at a configuration where both inner minima approach zero but on different sides of the corner; if h remains nonnegative while the true signed distance is negative, the claimed b=0 safety margin fails. A direct test is to run Example 4 with the hexagon aligned exactly with the rotating door edge and check whether any vertex enters the door while h stays nonnegative.

Watch

Extended reading notes

Core claim

The central claim is that the function h in Eq. (36), built from the under-approximation phi in Eq. (31), is a control barrier function candidate whose superlevel set {h>=0} is contained in the true safe set, so the safety-filter controller (6) renders the system collision-free with respect to the polytope environment. The load-bearing identity is Theorem 2: if the vertex-based phi = max over environment polytopes of min over their walls of min over agent vertices of the wall barriers is nonnegative, then every convex combination of the agent's vertices, i.e., the agent's convex hull, satisfies all the wall constraints, hence lies in the environment. The paper further claims, via the log-sum-exp approximations and a buffer b, that the smooth h under-approximates phi for suitably chosen b, which is what turns the geometric containment condition into a differentiable CBF condition; simulation examples carry the buffer to zero on the premise that phi's conservatism already supplies the margin.

Load-bearing premise

The paper's formal guarantee rests on the unverified premise that the smooth formula (36) is everywhere an under-approximation of the true signed distance, including when the buffer is set to zero, and that the resulting h satisfies the control barrier function condition (4); no explicit error bound or proof is given for the b=0 case.

Editorial extensions

If this is right

Systems with single-integrator dynamics can be made safe by filtering any desired velocity through the closed-form constraint (5) using h from (36), covering non-convex and time-varying polytope environments.
For second-order robotic models, the same closed-form h extends to a valid CBF via backstepping, so the method carries to manipulators, legged, wheeled, and flying robots.
Because h and its gradient are sums of exponentials of linear functions, the safe controller can be evaluated in real time on embedded hardware without online optimization.
The method is a safety filter, not a planner: a single goal waypoint can cause deadlock, but any nominal motion plan can be followed safely online.
The smoothing parameter kappa and buffer b trade off conservativeness and gradient steepness; increasing kappa sharpens the safe region at the cost of larger gradients.

Reading between the lines

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

If a formal error bound for the log-sum-exp outer max in nested max-min settings were derived, the b=0 cases would become certified; until then, controllers for those examples are validated by simulation rather than by the theorem.
The vertex-based under-approximation degrades as the number of convex polytopes grows, so approximating curved agents and obstacles by many small polytopes may demand larger buffers; an alternative using edges or faces would be less conservative but lose the closed form.
The same construction could be adapted to perception pipelines by treating environment barriers as time-varying parameters, since evaluating h and its gradient costs only evaluating linear forms and exponentials.
A natural testable extension is to measure how often the safety filter must override the nominal controller as kappa and b change, to calibrate conservativeness in practice.

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.

Desk Editor's Note

read the letter

Colleague,

The genuinely new thing here is a closed-form CBF candidate for polytope agents in polytope environments: Eq. (36), built from vertex-wise half-plane checks and log-sum-exp smoothing. That's a real simplification over existing optimization-based and cone-based polytope CBF methods, and Theorem 2 correctly proves the geometry: if the vertex-based phi in (31) is nonnegative, the agent's convex hull sits in the environment. The paper is clear, the examples (2D/3D, time-varying door, public code) are useful.

The soft spot is load-bearing for the "formal safety guarantees" claim. The paper suggests b=0 in (36) is enough because phi itself is conservative. That's not true. The outer log-sum-exp over-approximates max, and the conservatism of phi does not survive smoothing. Concrete counterexample: P = {x>=0} U {y>=0}, point agent at (-eps,-eps). phi = -eps < 0 (unsafe), but h with b=0 equals -eps + ln(2)/kappa, positive for small eps. So the superlevel set of h contains an unsafe state. The b=0 examples (3,4,5) are therefore not covered by the stated guarantee. This is fixable: choose b from the log-sum-exp error bound, e.g. b >= ln(N_p)/kappa or the sharper bound in [23, Thm. 5], and state it.

Second, the paper never verifies h satisfies Definition 1 as a CBF. For a single integrator this is automatic if grad h != 0 on the boundary of {h>=0}; that needs stating and checking. Minor compared to the buffer issue, but it should be closed to support the abstract.

The heavy reliance on the author's own [23] is not a problem; that composition result is general and independently useful.

Overall: the construction and geometric theorem are solid, the gap is specific, not fundamental. I'd send it to review — a competent referee can push for the buffer bound and the CBF regularity condition. I'd want the b=0 remark replaced with a principled buffer selection before trusting the formal claim. If I needed a polytope CBF, I'd cite the closed-form construction after that fix.

Recommendation: revise with explicit buffer bound and CBF regularity check; then accept.

Referee Report

3 major / 4 minor

Summary. The paper proposes a systematic method to construct smooth control barrier function (CBF) candidates in closed form for collision-free navigation of polytope-shaped agents in polytope-shaped environments. The safe set is expressed through max-min compositions of affine constraints over agent vertices, and the nonsmooth vertex-based function φ in Eq. (31) is proved (Theorem 2) to under-approximate the true safety condition: φ(x)≥0 implies that the convex hull of the agent is contained in the environment. The paper then proposes the log-sum-exp smoothed function h in Eq. (36) as a CBF candidate, with a buffer parameter b, and demonstrates the approach in 2D and 3D simulations, including time-varying environments, using b=0 in several examples. The abstract and conclusion claim that the proposed candidates lead to controllers with formal safety guarantees.

Significance. If the claims are established, the closed-form CBF candidate would be a practically attractive alternative to optimization-based polytope safety filters, enabling real-time implementation and easy extension to more complex robot models. Theorem 2 is a correct and clean geometric result, and the paper provides reproducible simulation code. However, the central formal guarantee is not established for the smoothed candidate: the paper does not prove that h under-approximates the safe set, and the b=0 choice used in the examples is demonstrably unsafe in simple configurations. These issues undermine the paper's main claim as currently stated, although they are potentially fixable with a derived buffer bound and a direct verification of the CBF condition.

major comments (3)

[Section III.B.3, Eq. (36)] The paper does not ensure that the smoothed function h in Eq. (36) is an under-approximation of the vertex-based function φ in Eq. (31). Consequently, the superlevel set C={x:h(x)≥0} may contain states that are not safe. The remark after Eq. (36) that b=0 'may be a sufficient buffer' because (31) is conservative is not a proof. A concrete counterexample: let the environment be P={p∈R^2 : p_1≥0 or p_2≥0}, represented as in (19) with J={1,2}, I_1={1}, I_2={2}, ψ_1(p)=p_1, ψ_2(p)=p_2, and take a point agent at p=(-ε,-ε). Then φ(p)=max{-ε,-ε}=-ε<0, so p is outside P. For b=0, Eq. (36) gives h(p)=(1/κ)ln(e^{κψ_1(p)}+e^{κψ_2(p)})=(1/κ)ln(2e^{-κε})=-ε+(ln 2)/κ, which is positive for any ε<(ln 2)/κ. Thus h's superlevel set contains an unsafe state, and the formal safety guarantee claimed in the abstract does not follow. The paper must either provide an explicit lower bound on b (e.g., derived from an approximation-error bound for the nested log-sum-exp composition) and use it in the examples, or adopt a smoothing that genuinely under-approximates the max function.
[Section III.B.3 and Definition 1] It is not shown that the smooth candidate h in Eq. (36) is a valid control barrier function for the single-integrator dynamics (12). The paper defines the safe set via h and invokes Theorem 1 to claim safety, but never verifies the CBF condition (4), i.e., that there exists α∈K_e such that sup_{u} h_dot(x,u) > -α(h(x)) for all x∈C. In particular, if ∇h(x)=0 for some x∈C, condition (4) can fail; the paper provides no argument that this cannot occur for the proposed h. A formal guarantee requires either a direct verification of (4) for the polytope environments under consideration or a reference to a known result that covers this case.
[Abstract and Section IV] The abstract and the conclusion state that the proposed approach yields controllers with 'formal guarantees' of collision-free motion. This overstates what is actually proven. The only theorem in the paper, Theorem 2, concerns the nonsmooth function φ in Eq. (31), not the smooth CBF candidate h in Eq. (36) that is used in the controller (6). Since the containment property h≤φ (or a suitable buffer-bound variant) is not established, and the CBF condition for h is not verified, the formal guarantee is not currently supported. The claims should be revised to match the proven results, or the missing analysis should be supplied.

minor comments (4)

[Section III.B.3, proof of Theorem 2] The step 'It can be shown that this ultimately yields' is terse; a one-line justification using monotonicity of min and max with respect to the pointwise inequalities in (34) would make the proof fully transparent.
[Eq. (36)] The notation with reciprocal sums in Eq. (36) is easy to misread. Consider defining the inner sum A_j = Σ_{i∈I_j} Σ_{k∈K} e^{-κ ψ_i(p_k)} and writing h = (1/κ) ln( Σ_j A_j^{-1} ) - b/κ, with a short explanation of the max-min composition.
[Example 4] For the time-varying revolving-door example, the paper should state explicitly that the index sets I_j and J remain constant as the door rotates, so that the composition (19) is time-independent as required by Remark 4.
[Section III.A.3] The paper mentions that b can be chosen based on the approximation error bound in [23, Thm. 5], but it does not carry out this calculation for the nested composition in Eq. (20) or Eq. (36). Even a conservative explicit bound for the simple convex-polytope case would strengthen the practical guidance.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the vertex-based safety theorem is proven from convexity, and the smoothed CBF is proposed as an approximation rather than being derived from the claimed result.

full rationale

The paper's central derivation is Theorem 2, which proves that if the exact vertex-based function phi in Eq. (31) is nonnegative, then the convex hull of the polytope agent is contained in the polytope environment. The proof is self-contained: it uses convexity of the barrier functions to show Eq. (33)-(35), and does not assume the conclusion. The smoothed CBF candidate h in Eq. (36) is explicitly presented as a log-sum-exp approximation of phi, not as a result that is definitionally tied to the safety claim. The paper's abstract claim of formal safety guarantees depends on h under-approximating the true safe set; this is a correctness question about whether the smoothing error is adequately bounded, especially for b=0 in Examples 3-5, but it is not a circular reduction of the target claim into its own input. The cited prior work, including the author's own [23], provides standard log-sum-exp approximation formulas and error bounds; these citations are used as tools, not as the sole justification that h is safe. No fitted parameter is renamed as a prediction, no uniqueness theorem is imported from the authors, and no ansatz is smuggled in solely via self-citation. The b=0 sufficiency remark is an unproven assertion and a potential correctness gap, but it does not make the derivation circular. Therefore the circularity score is 0.

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

No new physical entities are introduced. The free parameters kappa and b are design choices that the safety guarantee depends on. The largest hidden load is the unverified CBF property and the unproven under-approximation property for b=0.

free parameters (4)

smoothing parameter kappa = kappa = 5 in all examples
Controls the accuracy of the log-sum-exp approximation of min and max. Larger kappa reduces conservativeness but steepens the gradient. Chosen by hand, not fitted to data.
buffer parameter b = b = 0.7 in Example 2, b = 0 in Examples 3, 4, and 5
Intended to make h an under-approximation of the safety constraint. The b=0 choice in most examples is not justified with a formal error bound.
extended class-K function alpha = alpha(h) = 2h in all examples
Appears in the safety constraint in Eq. (5). Chosen by hand; any valid extended class-K function would work in theory.
desired controller gain K_p and saturation u_max = K_p = 1, u_max = 1 m/s
Used in the nominal goal-reaching controller in Eq. (26). These affect the trajectory and goal convergence but are not part of the safety certificate.

assumptions (4)

standard math Standard CBF theorem: if h is a control barrier function and the controller satisfies Eq. (5), then the closed-loop system is safe with respect to C.
Quoted as Theorem 1 from [1]. Accepted background in the CBF literature.
standard math Log-sum-exp formulas in Eq. (10) approximate max and min with controllable approximation error.
Taken from [23] and [24]. Used to smooth the nonsmooth max-min expression in Eq. (31) into the closed-form h in Eq. (36).
ad hoc to paper The smooth function h in Eq. (36) is a valid control barrier function for the single-integrator system.
The paper calls h a CBF candidate and applies Theorem 1 through the QP in Eq. (6), but it never proves Definition 1 or the needed gradient and feasibility conditions. This is a load-bearing unproven assumption.
domain assumption Time-varying environments have a fixed combinatorial topology, meaning the sets I_j and J do not change with time.
Remark 4 explicitly restricts the time-dependent extension to barriers with time-independent composition. The revolving door example relies on this assumption.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Navigating Polytopes with Safety: A Control Barrier Function Approach." pith.science (2026). https://pith.science/paper/Y2P3ZRJE

@misc{pith2026250517270,
  author       = {Pith},
  title        = {Pith review of: Navigating Polytopes with Safety: A Control Barrier Function Approach},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/Y2P3ZRJE}},
  note         = {Machine review of arXiv:2505.17270}
}

read the original abstract

Collision-free motion is a fundamental requirement for many autonomous systems. This paper develops a safety-critical control approach for the collision-free navigation of polytope-shaped agents in polytope-shaped environments. A systematic method is proposed to generate control barrier function candidates in closed form that lead to controllers with formal safety guarantees. The proposed approach is demonstrated through simulation, with obstacle avoidance examples in 2D and 3D, including dynamically changing environments.

Figures

Figures reproduced from arXiv: 2505.17270 by the authors.

**Figure 2.** Safety-critical navigation of a point agent around a polytope [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗

**Figure 3.** Safety-critical navigation of an ellipse agent around an ellipse [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗

Figures from the paper (2 more)

**Figure 4.** Figure 4: Safety-critical navigation of a hexagon agent through a revolving [PITH_FULL_IMAGE:figures/full_fig_p005_4.png]

**Figure 5.** Figure 5: Safety-critical navigation of a cube agent around a pyramid obstacle [PITH_FULL_IMAGE:figures/full_fig_p006_5.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

28 extracted references · 18 canonical work pages

[23]

Composing control barrier functions for complex safety specifications,

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

2023
[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 Transactions on Automatic Control, vol. 62, no. 8, pp. 3861–3876, 2017

2017
[2]

Advances in the theory of control barrier func- tions: Addressing practical challenges in safe control synthesis for autonomous and robotic systems,

K. Garg, J. Usevitch, J. Breeden, M. Black, D. Agrawal, H. Parwana, and D. Panagou, “Advances in the theory of control barrier func- tions: Addressing practical challenges in safe control synthesis for autonomous and robotic systems,”Annual Reviews in Control, vol. 57, p. 100945, 2024

2024
[3]

Collision avoidance control of robot arm considering the shape of the target system,

R. Hamatani and H. Nakamura, “Collision avoidance control of robot arm considering the shape of the target system,” in59th Annual Conference of the Society of Instrument and Control Engineers of Japan, 2020, pp. 1311–1316

work page 2020
[4]

Safety embedded differential dynamic programming using discrete barrier states,

H. Almubarak, K. Stachowicz, N. Sadegh, and E. A. Theodorou, “Safety embedded differential dynamic programming using discrete barrier states,”IEEE Robotics and Automation Letters, vol. 7, no. 2, pp. 2755–2762, 2022

work page 2022
[5]

Distributed safe nav- igation of multi-agent systems using control barrier function-based controllers,

P. Mestres, C. Nieto-Granda, and J. Cort ´es, “Distributed safe nav- igation of multi-agent systems using control barrier function-based controllers,”IEEE Robotics and Automation Letters, vol. 9, no. 7, pp. 6760–6767, 2024

2024
[6]

Safety barrier functions for human-robot interaction with industrial manipu- lators,

C. T. Landi, F. Ferraguti, S. Costi, M. Bonf `e, and C. Secchi, “Safety barrier functions for human-robot interaction with industrial manipu- lators,” in18th European Control Conference, 2019, pp. 2565–2570

work page 2019
[7]

Safety-critical manipulation for collision-free food prepara- tion,

A. W. Singletary, W. Guffey, T. Molnar, R. Sinnet, and A. D. Ames, “Safety-critical manipulation for collision-free food prepara- tion,”IEEE Robotics and Automation Letters, vol. 7, no. 4, pp. 10 954– 10 961, 2022

work page 2022

Show all 28 references

[8]

Synthesis of control barrier functions using a supervised machine learning approach,

M. Srinivasan, A. Dabholkar, S. Coogan, and P. A. Vela, “Synthesis of control barrier functions using a supervised machine learning approach,” inIEEE/RSJ International Conference on Intelligent Robots and Systems, 2020, pp. 7139–7145

2020
[9]

Learning barrier functions with memory for robust safe navigation,

K. Long, C. Qian, J. Cort ´es, and N. Atanasov, “Learning barrier functions with memory for robust safe navigation,”IEEE Robotics and Automation Letters, vol. 6, no. 3, pp. 4931–4938, 2021

2021
[10]

Safe robot navigation in a crowd combining NMPC and control barrier functions,

V . Vulcano, S. G. Tarantos, P. Ferrari, and G. Oriolo, “Safe robot navigation in a crowd combining NMPC and control barrier functions,” in61st IEEE Conference on Decision and Control, 2022, pp. 3321– 3328

2022
[11]

Safe control using vision-based control barrier function (V-CBF),

H. Abdi, G. Raja, and R. Ghabcheloo, “Safe control using vision-based control barrier function (V-CBF),” inIEEE International Conference on Robotics and Automation, 2023, pp. 782–788

2023
[12]

LiDAR-based on- line control barrier function synthesis for safe navigation in unknown environments,

S. Keyumarsi, M. W. S. Atman, and A. Gusrialdi, “LiDAR-based on- line control barrier function synthesis for safe navigation in unknown environments,”IEEE Robotics and Automation Letters, vol. 9, no. 2, pp. 1043–1050, 2024

2024
[13]

Point cloud-based control barrier function regression for safe and efficient vision-based control,

M. De Sa, P. Kotaru, and K. Sreenath, “Point cloud-based control barrier function regression for safe and efficient vision-based control,” inIEEE International Conference on Robotics and Automation, 2024, pp. 366–372

2024
[14]

Sailing through point clouds: Safe navigation using point cloud based control barrier functions,

B. Dai, R. Khorrambakht, P. Krishnamurthy, and F. Khorrami, “Sailing through point clouds: Safe navigation using point cloud based control barrier functions,”IEEE Robotics and Automation Letters, vol. 9, no. 9, pp. 7731–7738, 2024

2024
[15]

SAFER-Splat: A control barrier function for safe navigation with online Gaussian splatting maps,

T. Chen, A. Swann, J. Yu, O. Shorinwa, R. Murai, M. Kennedy III, and M. Schwager, “SAFER-Splat: A control barrier function for safe navigation with online Gaussian splatting maps,”arXiv preprint, no. arXiv:2409.09868, 2024

2024 arXiv
[16]

Neural control barrier functions for safe navigation,

M. Harms, M. Kulkarni, N. Khedekar, M. Jacquet, and K. Alexis, “Neural control barrier functions for safe navigation,” inIEEE/RSJ International Conference on Intelligent Robots and Systems, 2024, pp. 10 415–10 422

2024
[17]

Safe stabilizing control for polygonal robots in dynamic elliptical environments,

K. Long, K. Tran, M. Leok, and N. Atanasov, “Safe stabilizing control for polygonal robots in dynamic elliptical environments,” inAmerican Control Conference, 2024, pp. 312–317

2024
[18]

Safety-critical control and planning for obstacle avoidance between polytopes with control barrier functions,

A. Thirugnanam, J. Zeng, and K. Sreenath, “Safety-critical control and planning for obstacle avoidance between polytopes with control barrier functions,” inIEEE International Conference on Robotics and Automation, 2022, pp. 286–292

2022
[19]

Duality-based convex optimization for real-time obstacle avoid- ance between polytopes with control barrier functions,

——, “Duality-based convex optimization for real-time obstacle avoid- ance between polytopes with control barrier functions,” inAmerican Control Conference, 2022, pp. 2239–2246

2022
[20]

Polygonal cone control barrier functions (PolyC2BF) for safe navigation in cluttered environments,

M. Tayal and S. Kolathaya, “Polygonal cone control barrier functions (PolyC2BF) for safe navigation in cluttered environments,” inEuro- pean Control Conference, 2024, pp. 2212–2217

2024
[21]

Characterizing smooth safety filters via the implicit function theorem,

M. H. Cohen, P. Ong, G. Bahati, and A. D. Ames, “Characterizing smooth safety filters via the implicit function theorem,”IEEE Control Systems Letters, vol. 7, pp. 3890–3895, 2023

2023
[22]

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 Control Systems Letters, vol. 1, no. 2, pp. 310–315, 2017

2017
[24]

Control barrier functions for signal temporal logic tasks,

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

2019
[25]

Model-free safety-critical control for robotic systems,

T. G. Molnar, R. K. Cosner, A. W. Singletary, W. Ubellacker, and A. D. Ames, “Model-free safety-critical control for robotic systems,” IEEE Robotics and Automation Letters, vol. 7, no. 2, pp. 944–951, 2022

2022
[26]

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

2024
[27]

Safe back- stepping with control barrier functions,

A. J. Taylor, P. Ong, T. G. Molnar, and A. D. Ames, “Safe back- stepping with control barrier functions,” in61st IEEE Conference on Decision and Control, 2022, pp. 5775–5782

2022
[28]

Constructive safety- critical control: Synthesizing control barrier functions for partially feedback linearizable systems,

M. H. Cohen, R. K. Cosner, and A. D. Ames, “Constructive safety- critical control: Synthesizing control barrier functions for partially feedback linearizable systems,”IEEE Control Systems Letters, vol. 8, pp. 2229–2234, 2024

2024

Pith tools

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

[1] [23]

Composing control barrier functions for complex safety specifications,

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

2023

[2] [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 Transactions on Automatic Control, vol. 62, no. 8, pp. 3861–3876, 2017

2017

[3] [2]

Advances in the theory of control barrier func- tions: Addressing practical challenges in safe control synthesis for autonomous and robotic systems,

K. Garg, J. Usevitch, J. Breeden, M. Black, D. Agrawal, H. Parwana, and D. Panagou, “Advances in the theory of control barrier func- tions: Addressing practical challenges in safe control synthesis for autonomous and robotic systems,”Annual Reviews in Control, vol. 57, p. 100945, 2024

2024

[4] [3]

Collision avoidance control of robot arm considering the shape of the target system,

R. Hamatani and H. Nakamura, “Collision avoidance control of robot arm considering the shape of the target system,” in59th Annual Conference of the Society of Instrument and Control Engineers of Japan, 2020, pp. 1311–1316

work page 2020

[5] [4]

Safety embedded differential dynamic programming using discrete barrier states,

H. Almubarak, K. Stachowicz, N. Sadegh, and E. A. Theodorou, “Safety embedded differential dynamic programming using discrete barrier states,”IEEE Robotics and Automation Letters, vol. 7, no. 2, pp. 2755–2762, 2022

work page 2022

[6] [5]

Distributed safe nav- igation of multi-agent systems using control barrier function-based controllers,

P. Mestres, C. Nieto-Granda, and J. Cort ´es, “Distributed safe nav- igation of multi-agent systems using control barrier function-based controllers,”IEEE Robotics and Automation Letters, vol. 9, no. 7, pp. 6760–6767, 2024

2024

[7] [6]

Safety barrier functions for human-robot interaction with industrial manipu- lators,

C. T. Landi, F. Ferraguti, S. Costi, M. Bonf `e, and C. Secchi, “Safety barrier functions for human-robot interaction with industrial manipu- lators,” in18th European Control Conference, 2019, pp. 2565–2570

work page 2019

[8] [7]

Safety-critical manipulation for collision-free food prepara- tion,

A. W. Singletary, W. Guffey, T. Molnar, R. Sinnet, and A. D. Ames, “Safety-critical manipulation for collision-free food prepara- tion,”IEEE Robotics and Automation Letters, vol. 7, no. 4, pp. 10 954– 10 961, 2022

work page 2022