Pith. sign in

REVIEW 3 major objections 4 minor 30 references

Opt-ODENet: A Neural ODE Framework with Differentiable QP Layers for Safe and Stable Control Design (longer version)

T0 review · 3 major / 4 minor · reviewed 2026-08-16 · deepseek-v4-flash

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

Pith's one-line read A Neural ODE controller can be trained through a differentiable QP layer that projects each command onto a barrier-function safety constraint, with a Lyapunov loss for stability.

desk verdict A genuinely useful integration of Neural ODEs with differentiable CBF-QP layers, but Proposition 4 omits the QP layer's state dependence, so the main gradient claim needs repair before the paper's guarantees can be taken at face value. read the letter →

arxiv 2504.17139 v1 pith:LKOONN7G submitted 2025-04-23 eess.SY cs.SY

Keyan Miao , Liqun Zhao , Han Wang , Konstantinos Gatsis , Antonis Papachristodoulou This is my paper · ORCID
classification eess.SYcs.SY
keywords NeuralODEdifferentiableQPlayercontrolbarrierfunctionLyapunovsafeadjointmethodhigher-orderCBFoptimal
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

Opt-ODENet claims that a state-feedback controller can be learned end-to-end by passing the network's raw command through a differentiable quadratic-programming (QP) layer that projects the command onto a control barrier function (CBF) constraint — a scalar safety certificate whose positivity defines the set of safe states — so that safety is enforced as a hard requirement. A control Lyapunov function (CLF, a scalar certificate whose decrease implies convergence) is placed in the loss, and the paper proves that gradients flow backward through both the Neural ODE dynamics and the QP solution by combining the adjoint method with differentiation of the QP's KKT system; this makes the CBF's class-$\mathcal{K}$ coefficient learnable. If correct, the framework designs safe, stabilizing controllers without a nominal controller, reference trajectories, or large datasets. Experiments on a unicycle obstacle-avoidance task and a higher-order-CBF car-following task support the claim.

What carries the argument

The load-bearing object is the differentiable CBF-QP projection layer: a quadratic program whose objective is $\tfrac{1}{2}\|u-u_{nn}\|^2$ with $u_{nn}=\pi(x,\theta_1)$, and whose inequality constraint is the CBF condition $\dot{B}(x,u)+\alpha(B(x),\theta_2)\ge 0$. The layer is differentiable because its backward pass differentiates the KKT conditions (15)-(20), producing the sensitivities $\partial u^*/\partial q$ and $\partial u^*/\partial h$ that the adjoint equations (11) need for $\nabla_{\theta_1}\ell$ and $\nabla_{\theta_2}\ell$. The second mechanism is the CLF-based Lyapunov loss $\ell=\mathbb{E}\left[\int_{t_0}^{t_f} \max\{0,\dot{V}(x,u)+\gamma V(x)\}\,dt\right]$; when training drives this loss to zero, the closed-loop system inherits exponential convergence of the form $\Phi(x(t))\le \Phi(x(t_0))e^{-\kappa t}$.

What would settle it

Run the trained controller from an initial state that approaches the boundary of the safe set while the raw network command pushes inward; if at any reached state the constraint set $\{u : \dot{B}(x,u)+\alpha(B(x))\ge 0\}$ is empty, the QP has no feasible output and the claimed safety-by-construction guarantee fails at that point.

Watch

Extended reading notes

Core claim

On the paper's own terms, the central claim is that for a known control-affine system $\dot{x}=f(x)+g(x)u$ and a known CBF $B$, the policy $u=\tau(x,\theta_1,\theta_2)$ obtained by solving the QP $\min_u \tfrac{1}{2}\|u-\pi(x,\theta_1)\|^2$ subject to $\frac{\partial B}{\partial x}f(x)+\frac{\partial B}{\partial x}g(x)u+\alpha(B(x),\theta_2)\ge 0$ can be trained directly by backpropagating a CLF-based loss through the Neural ODE and through the QP. Proposition 4 establishes the adjoint equations (11) for the parameter gradients, with the QP sensitivity entering through the derivative of the KKT system (equations (15)-(20)). The learned class-$\mathcal{K}$ coefficient $\kappa=13.38$ yields collision-free unicycle trajectories with mean error $0.2036$, the best among the tested safe configurations in Table 1, and in the car-following HOCBF setting the learned per-constraint $\kappa$ gives zero safety violations and better reward than fixed choices.

Load-bearing premise

The load-bearing premise is that the CBF-QP layer is feasible and differentiable at every state visited during training and deployment — the paper's own appendix records a car-following case where infeasibility made training problematic — and, for the exponential-convergence statement, that the Lyapunov loss can actually be driven to zero.

Editorial extensions

If this is right

  • Deployment can use the trained policy as the final controller: every command is the QP projection of the neural output, so no separate runtime safety filter is needed as long as the QP remains feasible.
  • Because the CBF class-$\mathcal{K}$ coefficient is learned, the user does not need to hand-tune how strongly safety intervenes; the optimizer finds a value such as $\kappa = 13.38$ that balances convergence against conservatism.
  • The approach removes the need for a nominal safe controller, reference trajectories, or large datasets: training solves the optimal control problem directly from sampled initial states for known dynamics.
  • If the Lyapunov loss reaches zero, the learned closed-loop system converges exponentially at rate $\kappa$, so the same certificate used in training doubles as a stability guarantee.
  • Higher-order CBFs extend the framework to constraints with relative degree greater than one, such as car following under acceleration control.

Reading between the lines

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

  • The paper leaves implicit that the same adjoint-plus-KKT derivation would extend to other differentiable optimization layers (for example, a second-order cone or nonlinear projection), broadening the safety mechanism beyond quadratic programs without changing the training scheme.
  • Because the unicycle case learns only a scalar class-$\mathcal{K}$ coefficient while the car-following case needs per-constraint coefficients, a natural extension is to make $\alpha$ state- or constraint-dependent; the multi-constraint result already hints that a single scalar is a bottleneck.
  • One could test generalization by training in one obstacle layout and deploying in a different one; if the learned $\kappa$ transfers, it would suggest the QP projection, not memorized geometry, is carrying the safety guarantee.
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

3 major / 4 minor

Summary. The paper proposes Opt-ODENet, a Neural ODE-based control framework in which a Control Barrier Function (CBF) constraint is enforced as a hard requirement through a differentiable Quadratic Programming (QP) layer, while a Control Lyapunov Function (CLF) is used as a training loss to promote convergence to a target. The main theoretical claim is Proposition 4: the loss gradients with respect to the controller network parameters θ1 and the CBF class-K parameters θ2 can be computed by solving an adjoint system whose right-hand side involves, among other terms, the derivative of the QP layer solution τ with respect to its inputs, obtained via KKT sensitivity analysis. The paper presents a unicycle obstacle-avoidance experiment and a multi-car HOCBF experiment, reporting that learning the class-K coefficient κ (13.38 for the unicycle; {18.6, 96.04} for the cars) yields the best safety-performance trade-off among the tested configurations. The long version includes an appendix with additional derivations and experimental details.

Significance. If the central gradient derivation were correct, the paper would contribute a useful design pattern: combining Neural ODEs with differentiable optimization layers to train state-feedback controllers that are safe by construction and whose stability is encouraged by a CLF term. The idea of learning the CBF class-K parameter through gradient-based training rather than fixing it a priori is interesting and is supported by the empirical comparisons in Tables 1. The paper also provides a public code link, which is commendable for reproducibility. However, the load-bearing theoretical claim (Proposition 4) contains a chain-rule gap that undermines the stated adjoint computation, and the hard-safety guarantee is qualified by the documented infeasibility of the QP layer during training. These issues prevent the paper from currently establishing its central claims, although they appear to be locally fixable.

major comments (3)
  1. [Section 4, Eq. (13)] The chain rule for ∂τ/∂x in Eq. (13) is incomplete. The CBF-QP layer (9) has G = -∂B/∂x g(x) and h = ∂B/∂x f(x) + α(B(x), θ2), both of which depend on the state x. The full derivative of the QP solution with respect to x is ∂τ/∂x = (∂τ/∂π)(∂π/∂x) + (∂τ/∂G)(∂G/∂x) + (∂τ/∂h)(∂h/∂x), where ∂τ/∂G and ∂τ/∂h are precisely the KKT sensitivities (20). Eq. (13) omits the last two terms, so the adjoint state p in (11) does not satisfy the correct adjoint equation and μ1(t0), μ2(t0) are not the gradients of the loss ℓ. This is an internal inconsistency in the proof of Proposition 4, not merely a stylistic issue. If the actual implementation uses unrolled automatic differentiation through the ODE solver and the QP layer, then the paper should state that Proposition 4 does not describe the method being evaluated, or should supply the corrected adjoint equations.
  2. [Section 4, Eq. (9); Appendix 7.2.2] The paper assumes the CBF-QP layer is feasible and differentiable at every state visited during training, but this is not guaranteed. The appendix (Section 7.2.2) documents that in the car-following HOCBF case with κ={2,1}, 'certain infeasibility happens, which makes the whole training problematic.' Since the forward pass of the QP layer fails when no feasible control exists, the abstract's statement that CBFs 'enforce real-time safety' as hard requirements is not supported by the presented method. The paper should either provide a sufficient condition for feasibility, add a relaxation (such as slack variables) with an explicit safety guarantee, or clearly bound the set of states for which the QP is feasible.
  3. [Abstract; Remark 5] The abstract claims that 'Stability and convergence are ensured through Control Lyapunov Functions (CLFs) in the loss function'. However, Remark 5 states that the convergence expression (26) holds only if there exists a parameter θ* satisfying ℓ(θ*)=0. Gradient-based training provides no guarantee that such a parameter is reached, and the experiments only show empirical convergence losses. The claims in the abstract and Section 6 should be rephrased to reflect the conditional nature of the stability property, e.g., 'encouraged' or 'promoted' rather than 'ensured'.
minor comments (4)
  1. [Figure 2 caption] The caption says 'Distance over time with varying ℓ', but the text refers to varying γ (the CLF decay rate); the caption should be corrected to avoid ambiguity.
  2. [Appendix 7.1.2, Eq. (51)] The expression for ∂ℓ/∂G appears to have a typographical inconsistency: the term 'λ∗d⊤u' is ambiguous (the standard OptNet result is λ* du^T plus a term involving D(λ*)dλ z*^T); please clarify the notation.
  3. [Throughout the manuscript] There are many rendering artifacts in the equations (e.g., '⁄𝑥' instead of \dot{x}, '𝜃' for θ, and fragmented over-dots), which make the paper difficult to read. The authors should ensure the final version is typeset cleanly.
  4. [Section 7.1.1, Eq. (32)] The adjoint equation for p in the appendix writes ∂L/∂z in the p dynamics, but the running cost L depends on x, t, and θ; this should be ∂L/∂x for consistency with the main text.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the adjoint and QP-sensitivity derivations are standard, externally grounded results, and the learned kappa is reported as a fitted parameter, not as a prediction.

full rationale

The claimed derivation chain is self-contained against external benchmarks. The adjoint system (11) is the textbook Pontryagin adjoint method applied to the Neural ODE formulation in (1)-(2), and the QP sensitivity in equations (20) and (51) is the standard OptNet KKT-Jacobian result, re-derived here via the implicit function theorem in Theorem 6 with external references (Dontchev and Rockafellar 2009; Amos and Kolter 2017). The learned class-K coefficient kappa = 13.38 is a fitted parameter, reported as a learned value rather than as a prediction of an independent quantity, so Table 1's comparison against fixed kappa values is an ablation, not a circular prediction. The CLF-based stability statement is explicitly conditional: Remark 5 guarantees exponential convergence only if a parameter theta* with l(theta*)=0 exists, and the safety property is enforced by construction through the CBF-QP layer (9), whose constraints are exactly the CBF condition from Definition 3. No load-bearing step reduces to a self-citation chain; the authors' prior works are cited only for background. The appendix's admission of HOCBF infeasibility in certain kappa regimes (Appendix 7.2.2) and any omitted state-dependence in equation (13) are technical or correctness concerns, not circular reductions, because they do not make the output equal to an input by definition.

Assumptions & free parameters 4 free parameters · 5 assumptions · 0 invented entities

The central claim rests on five premises: known dynamics f, g; a known a priori CBF; feasibility and differentiability of the QP along all trajectories; successful minimization of the CLF loss; and standard implicit-function conditions. The first two are stated, the third is assumed in the main text and shown to fail in the appendix, and the fourth is an unverified training success condition that turns Remark 5's conditional guarantee into the abstract's unconditional claim of stability.

free parameters (4)
  • kappa (CBF class-K linear coefficient, unicycle) = 13.38 (learned), compared with fixed 5 and 10
    Learned during training to set the aggressiveness of the safety constraint in the QP layer (Section 5, Table 1). It is the main tuning knob of the paper and is fitted to the training objective.
  • kappa (HOCBF class-K coefficients, cars) = {2,1}, {60,900}, {18.6, 96.04} (learned)
    Per-constraint class-K coefficients for the two car-following CBFs; learned values are reported and compared with fixed choices (Appendix 7.2.2).
  • gamma (CLF decay rate in Lyapunov loss) = not reported numerically; varied as clf-10/20/30 in Figure 2
    The class-K rate in the CLF condition (4) and loss (24)-(25), chosen by hand to set the convergence rate; no automatic tuning.
  • CBF/CLF geometry constants (delta1, delta2, l_p) = not reported numerically
    Hand-chosen safety distance, target radius, and lookahead distance defining the CBF h(x) and CLF V(x) in Section 5; they determine the problem and are not learned.
assumptions (5)
  • domain assumption The drift f(x) and control matrix g(x) of the control-affine system (7) are known exactly and time-invariant.
    Section 3 states the framework assumes known f, g; the whole optimization (8) depends on them.
  • domain assumption A valid CBF B(x) defining the safe set is known a priori.
    Section 3 assumes a CBF is known, claiming this follows from f, g being known, which is not logically implied; the CBF must be designed separately.
  • domain assumption The CBF-QP (9) is feasible and its KKT Jacobian (18) is invertible at every state along the trajectory.
    Required for the forward pass to exist and for the gradient formulas (15)-(20) to hold; the appendix documents infeasibility in the car experiment (Appendix 7.2.2).
  • ad hoc to paper Gradient descent reaches a parameter with (near-)zero CLF loss (25).
    Remark 5's exponential convergence guarantee is conditional on the existence of theta* with loss(theta*) = 0; the paper never proves training reaches such a point, so the abstract's 'ensured' stability rests on this unverified premise.
  • standard math Implicit Function Theorem conditions for QP differentiation: Q positive definite, A and G full rank.
    Theorem 6 in Appendix 7.1.2 provides the implicit function theorem used to differentiate the QP solution; the assumptions are standard but not checked for the specific experiments.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Opt-ODENet: A Neural ODE Framework with Differentiable QP Layers for Safe and Stable Control Design (longer version)." pith.science (2026). https://pith.science/paper/LKOONN7G

@misc{pith2026250417139,
  author       = {Pith},
  title        = {Pith review of: Opt-ODENet: A Neural ODE Framework with Differentiable QP Layers for Safe and Stable Control Design (longer version)},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/LKOONN7G}},
  note         = {Machine review of arXiv:2504.17139}
}
abstract

Designing controllers that achieve task objectives while ensuring safety is a key challenge in control systems. This work introduces Opt-ODENet, a Neural ODE framework with a differentiable Quadratic Programming (QP) optimization layer to enforce constraints as hard requirements. Eliminating the reliance on nominal controllers or large datasets, our framework solves the optimal control problem directly using Neural ODEs. Stability and convergence are ensured through Control Lyapunov Functions (CLFs) in the loss function, while Control Barrier Functions (CBFs) embedded in the QP layer enforce real-time safety. By integrating the differentiable QP layer with Neural ODEs, we demonstrate compatibility with the adjoint method for gradient computation, enabling the learning of the CBF class-$\mathcal{K}$ function and control network parameters. Experiments validate its effectiveness in balancing safety and performance.

Figures

Figures reproduced from arXiv: 2504.17139 by the authors.

Figure 1
Figure 1. Schematics of the Neural ODE-based controller with a differentiable CBF-QP layer en￾forcing safety constraints The differential equations (11) are derived by applying the adjoint method to an optimal control problem. In the following proof, we demonstrate the gradient computation for Neural ODEs with an embedded differentiable QP layer, leveraging the interconnected structure of the controller network. Due to space … view at source ↗
Figure 2
Figure 2. Distance over time with varying ℓ 0.0 0.2 0.4 0.6 0.8 1.0 t 0.0 0.2 0.4 0.6 0.8 1.0 Distance to the target fixed = 5 fixed = 10 learned = 13.38 [PITH_FULL_IMAGE:figures/full_fig_p009_2.png] view at source ↗
Figure 4
Figure 4. Tested Trajectories under Different CBF-QP Settings [PITH_FULL_IMAGE:figures/full_fig_p009_4.png] view at source ↗
Figures from the paper (3 more)
Figure 5
Figure 5. Figure 5: Training results on Simulated Cars with HOCBF [PITH_FULL_IMAGE:figures/full_fig_p010_5.png]
Figure 6
Figure 6. Figure 6: Test Trajectories under Different HOCBF-QP Settings for Unicycle [PITH_FULL_IMAGE:figures/full_fig_p018_6.png]
Figure 7
Figure 7. Figure 7: Training results on Simulated Cars under Different HOCBF-QP Settings [PITH_FULL_IMAGE:figures/full_fig_p019_7.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

30 extracted references · 20 canonical work pages

  1. [1]

    Rapidly exponentially stabilizing control lyapunov functions and hybrid zero dynamics

    Aaron D Ames, Kevin Galloway, Koushil Sreenath, and Jessy W Grizzle. Rapidly exponentially stabilizing control lyapunov functions and hybrid zero dynamics. IEEE Transactions on Automatic Control, 59 0 (4): 0 876--891, 2014

    work page 2014

  2. [2]

    Control barrier function based quadratic programs for safety critical systems

    Aaron D Ames, Xiangru Xu, Jessy W Grizzle, and Paulo Tabuada. Control barrier function based quadratic programs for safety critical systems. IEEE Transactions on Automatic Control, 62 0 (8): 0 3861--3876, 2016

    2016

  3. [3]

    Control barrier functions: Theory and applications

    Aaron D Ames, Samuel Coogan, Magnus Egerstedt, Gennaro Notomista, Koushil Sreenath, and Paulo Tabuada. Control barrier functions: Theory and applications. In 2019 18th European control conference (ECC), pages 3420--3431. IEEE, 2019

    2019

  4. [4]

    Optnet: Differentiable optimization as a layer in neural networks

    Brandon Amos and J Zico Kolter. Optnet: Differentiable optimization as a layer in neural networks. In International conference on machine learning, pages 136--145. PMLR, 2017

    2017

  5. [5]

    Neural odes for data-driven automatic self-design of finite-time output feedback control for unknown nonlinear dynamics

    Simon Bachhuber, Ive Weygers, and Thomas Seel. Neural odes for data-driven automatic self-design of finite-time output feedback control for unknown nonlinear dynamics. IEEE Control Systems Letters, 2023

    work page 2023

  6. [6]

    Ai pontryagin or how artificial neural networks learn to control dynamical systems

    Lucas B \"o ttcher, Nino Antulov-Fantulin, and Thomas Asikis. Ai pontryagin or how artificial neural networks learn to control dynamical systems. Nature communications, 13 0 (1): 0 333, 2022

    work page 2022

  7. [7]

    Neural ordinary differential equations

    Ricky TQ Chen, Yulia Rubanova, Jesse Bettencourt, and David K Duvenaud. Neural ordinary differential equations. Advances in neural information processing systems, 31, 2018

    2018

  8. [8]

    Safe nonlinear control using robust neural lyapunov-barrier functions

    Charles Dawson, Zengyi Qin, Sicun Gao, and Chuchu Fan. Safe nonlinear control using robust neural lyapunov-barrier functions. In Conference on Robot Learning, pages 1724--1735. PMLR, 2022

    2022

Show all 30 references
  1. [9]

    Neural networks with physics-informed architectures and constraints for dynamical systems modeling

    Franck Djeumou, Cyrus Neary, Eric Goubault, Sylvie Putot, and Ufuk Topcu. Neural networks with physics-informed architectures and constraints for dynamical systems modeling. In Learning for Dynamics and Control Conference, pages 263--277. PMLR, 2022

    2022

  2. [10]

    Implicit functions and solution mappings, volume 543

    Asen L Dontchev and R Tyrrell Rockafellar. Implicit functions and solution mappings, volume 543. Springer, 2009

    2009

  3. [11]

    Deep residual learning for image recognition

    Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun. Deep residual learning for image recognition. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 770--778, 2016

    2016

  4. [12]

    Lyapunov neural ode feedback control policies

    Joshua Hang Sai Ip, Georgios Makrygiorgos, and Ali Mesbah. Lyapunov neural ode feedback control policies. arXiv preprint arXiv:2409.00393, 2024

    2024

  5. [13]

    Neural certificates for safe control policies

    Wanxin Jin, Zhaoran Wang, Zhuoran Yang, and Shaoshuai Mou. Neural certificates for safe control policies. arXiv preprint arXiv:2006.08465, 2020

    2006 arXiv

  6. [14]

    Modeling trajectories with neural ordinary differential equations

    Yuxuan Liang, Kun Ouyang, Hanshu Yan, Yiwei Wang, Zekun Tong, and Roger Zimmermann. Modeling trajectories with neural ordinary differential equations. In IJCAI, pages 1498--1504, 2021

    2021

  7. [15]

    Learning robust state observers using neural odes

    Keyan Miao and Konstantinos Gatsis. Learning robust state observers using neural odes. In Learning for Dynamics and Control Conference, pages 208--219. PMLR, 2023

    2023

  8. [16]

    How deep do we need: Accelerating training and inference of neural ODE s via control perspective

    Keyan Miao and Konstantinos Gatsis. How deep do we need: Accelerating training and inference of neural ODE s via control perspective. In Proceedings of the 41st International Conference on Machine Learning, volume 235 of Proceedings of Machine Learning Research. PMLR, 21--27 Jul 2024

    2024

  9. [17]

    Learning complex motion plans using neural odes with safety and stability guarantees

    Farhad Nawaz, Tianyu Li, Nikolai Matni, and Nadia Figueroa. Learning complex motion plans using neural odes with safety and stability guarantees. In 2024 IEEE International Conference on Robotics and Automation (ICRA), pages 17216--17222, 2024. doi:10.1109/ICRA57147.2024.10611584

    2024

  10. [18]

    Safe optimal control using stochastic barrier functions and deep forward-backward sdes

    Marcus Pereira, Ziyi Wang, Ioannis Exarchos, and Evangelos Theodorou. Safe optimal control using stochastic barrier functions and deep forward-backward sdes. In Conference on Robot Learning, pages 1783--1801. PMLR, 2021

    2021

  11. [19]

    Mathematical theory of optimal processes

    LS Pontryagin. Mathematical theory of optimal processes. CRC press, 1987

    1987

  12. [20]

    Lyanet: A lyapunov framework for training neural odes

    Ivan Dario Jimenez Rodriguez, Aaron Ames, and Yisong Yue. Lyanet: A lyapunov framework for training neural odes. In International Conference on Machine Learning, pages 18687--18703. PMLR, 2022

    2022

  13. [21]

    Neural odes as feedback policies for nonlinear optimal control

    Ilya Orson Sandoval, Panagiotis Petsagkourakis, and Ehecatl Antonio del Rio-Chanona. Neural odes as feedback policies for nonlinear optimal control. IFAC-PapersOnLine, 56 0 (2): 0 4816--4821, 2023

    2023

  14. [22]

    A ‘universal’construction of artstein's theorem on nonlinear stabilization

    Eduardo D Sontag. A ‘universal’construction of artstein's theorem on nonlinear stabilization. Systems & control letters, 13 0 (2): 0 117--123, 1989

    1989

  15. [23]

    Learning for safety-critical control with control barrier functions

    Andrew Taylor, Andrew Singletary, Yisong Yue, and Aaron Ames. Learning for safety-critical control with control barrier functions. In Learning for Dynamics and Control, pages 708--717. PMLR, 2020

    2020

  16. [24]

    Safety verification and controller synthesis for systems with input constraints

    Han Wang, Kostas Margellos, and Antonis Papachristodoulou. Safety verification and controller synthesis for systems with input constraints. IFAC-PapersOnLine, 56 0 (2): 0 1698--1703, 2023

    2023

  17. [25]

    Convex co-design of control barrier function and safe feedback controller under input constraints

    Han Wang, Kostas Margellos, Antonis Papachristodoulou, and Claudio De Persis. Convex co-design of control barrier function and safe feedback controller under input constraints. arXiv preprint arXiv:2403.11763, 2024

    2024 arXiv

  18. [26]

    Control barrier functions for systems with high relative degree

    Wei Xiao and Calin Belta. Control barrier functions for systems with high relative degree. In 2019 IEEE 58th conference on decision and control (CDC), pages 474--479. IEEE, 2019

    2019

  19. [27]

    Safe neural control for non-affine control systems with differentiable control barrier functions

    Wei Xiao, Ross Allen, and Daniela Rus. Safe neural control for non-affine control systems with differentiable control barrier functions. In 2023 62nd IEEE Conference on Decision and Control (CDC), pages 3366--3371. IEEE, 2023 a

    2023

  20. [28]

    Barriernet: Differentiable control barrier functions for learning of safe robot control

    Wei Xiao, Tsun-Hsuan Wang, Ramin Hasani, Makram Chahine, Alexander Amini, Xiao Li, and Daniela Rus. Barriernet: Differentiable control barrier functions for learning of safe robot control. IEEE Transactions on Robotics, 39 0 (3): 0 2289--2307, 2023 b

    2023

  21. [29]

    Stable and safe reinforcement learning via a barrier-lyapunov actor-critic approach

    Liqun Zhao, Konstantinos Gatsis, and Antonis Papachristodoulou. Stable and safe reinforcement learning via a barrier-lyapunov actor-critic approach. In 2023 62nd IEEE Conference on Decision and Control (CDC), pages 1320--1325. IEEE, 2023

    2023

  22. [30]

    Nlbac: A neural ode-based algorithm for state-wise stable and safe reinforcement learning

    Liqun Zhao, Keyan Miao, Hongpeng Cao, Konstantinos Gatsis, and Antonis Papachristodoulou. Nlbac: A neural ode-based algorithm for state-wise stable and safe reinforcement learning. Neurocomputing, page 130041, 2025

    2025

Pith tools

Reviewed August 16, 2026 · model on record in the stance chip above.