arxiv: 2605.05575 · v1 · submitted 2026-05-07 · 📡 eess.SY · cs.RO· cs.SY· math.OC

Recognition: unknown

Maximal Controlled Invariant-MPC: Enhancing Feasibility and Reducing Conservatism through Terminal CBF Constraint in Safety-Critical Control

Tanmay Dokania , Yashwanth Kumar Nakka

Authors on Pith no claims yet

Pith reviewed 2026-05-08 07:31 UTC · model grok-4.3

classification 📡 eess.SY cs.ROcs.SYmath.OC

keywords Model Predictive ControlControl Barrier Functionssafety-critical controlterminal constraintsfeasibilitynonholonomic systems

0 comments

The pith

MPC formulation with Control Barrier Function as terminal constraint improves feasibility and enlarges reachable sets as prediction horizon increases.

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

The paper proposes using a Control Barrier Function directly as the terminal constraint in a Model Predictive Control setup for safety-critical systems. This is shown to reduce conservatism compared to standard approaches, leading to better feasibility of the optimization and larger sets of states from which the system can safely operate. Proofs are constructive, allowing the solver to be warm-started from previous solutions, which cuts computation time. Simulations on a nonholonomic vehicle demonstrate fewer infeasible starting points and the ability to follow paths that enter regions marked unsafe by the CBF.

Core claim

A Model Predictive Control formulation that incorporates a Control Barrier Function as its terminal constraint is proven to enhance feasibility and expand reachable sets as the prediction horizon lengthens. The constructive proofs enable warm-starting of the nonlinear program, substantially lowering solve times. In simulations, this reduces the number of infeasible points by factors of 1.7 to 2.7 and permits tracking of trajectories lying inside the CBF-defined unsafe region.

What carries the argument

The terminal Control Barrier Function constraint placed at the end of the MPC prediction horizon, which enforces safety and supports recursive feasibility without additional tuning.

If this is right

Feasibility of the MPC problem increases with longer prediction horizons.
The set of safe reachable states grows, allowing operation closer to the safety boundary.
Computational burden decreases through warm-starting of the optimization solver.
The closed-loop system can follow reference trajectories that would otherwise be blocked by conservative constraints.

Where Pith is reading between the lines

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

Applying this terminal constraint approach to higher-dimensional or uncertain systems could yield similar gains in performance.
Combining it with online CBF adaptation might further minimize conservatism in real-time settings.
Testing on physical hardware would confirm whether the numerical improvements translate to practical safety-critical applications.

Load-bearing premise

That a given Control Barrier Function can be imposed directly as the terminal constraint while maintaining recursive feasibility and safety guarantees without introducing new sources of conservatism.

What would settle it

A numerical example or system where extending the prediction horizon in this MPC-CBF setup either fails to increase the feasible region or reduces the reachable set size.

read the original abstract

Optimal control for safety-critical systems is often dependent on the conservativeness of constraints. Control Barrier Functions (CBFs) serve as a medium to represent such constraints, but constructing a minimally conservative CBF is a computationally intractable problem. Therefore, approaches that can guarantee safety while reducing conservatism will help improve the optimality of the system under consideration. Here, we present a Model Predictive Control (MPC) formulation using CBF as a terminal constraint, which is proven to improve feasibility and reachable sets with increasing prediction horizon. The constructive nature of the proofs allows for warm-starting the nonlinear optimization problem, thereby reducing the computational time substantially. Simulations are set up for a simple nonholonomic system to numerically validate the results, and it is observed that the number of infeasible points decreased by a factor of 1.7 to 2.7. The increase in reachable state space was demonstrated by the ability of the system to track trajectories that are entirely inside the unsafe region of the control barrier function.

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.

Desk Editor's Note private letter to a colleague

Terminal CBF as MPC constraint enlarges feasible sets with horizon length and cuts infeasible points 1.7-2.7x in nonholonomic tests, via constructive proofs that also support warm-starting.

read the letter

The core idea is an MPC formulation that places a control barrier function directly as the terminal constraint. The paper claims this yields larger feasible regions and reachable sets as the prediction horizon increases, with constructive proofs that also enable warm-starting the nonlinear program for faster solves. Simulations on a nonholonomic system back the numbers: infeasible points drop by factors of 1.7 to 2.7, and the closed-loop system can track trajectories that sit inside the CBF's unsafe region. That directly tackles the conservatism that often forces overly cautious behavior in safety-critical MPC. The work is clearest on the horizon-dependent improvement and the computational side benefit from the proof structure. The nonholonomic example is a clean test bed that shows the practical effect without extra machinery. Soft spots are modest. The central argument depends on the terminal CBF set remaining controlled-invariant under the MPC policy, and while the stress-test sees no inconsistency in the stated claims, the full derivations would need checking for hidden assumptions on the CBF or dynamics that could reintroduce conservatism. The simulation results report clear gains but would be stronger with explicit baseline comparisons to standard CBF-MPC or terminal-set methods and some measure of variance. Nothing in the abstract or stress-test points to circularity or fitted quantities. This paper is for control researchers and engineers working on safe MPC for robotics or vehicles, where feasibility limits are a daily constraint. A reader looking for concrete ways to loosen barrier conservatism while keeping guarantees would get usable material. It deserves peer review because the formulation is specific, the claims are testable, and the evidence combines proofs with reproducible simulation outcomes.

Referee Report

1 major / 2 minor

Summary. The manuscript proposes a Model Predictive Control (MPC) formulation that imposes a Control Barrier Function (CBF) as a terminal constraint, termed Maximal Controlled Invariant-MPC. It claims constructive proofs that feasibility and the size of the reachable set both improve with increasing prediction horizon, that the proofs enable warm-starting of the nonlinear program, and that numerical experiments on a nonholonomic system show a 1.7-2.7 factor reduction in infeasible points together with the ability to track trajectories lying inside the unsafe region of the CBF.

Significance. If the terminal-CBF construction is shown to be controlled-invariant under the MPC policy and to preserve recursive feasibility without hidden conservatism, the approach would meaningfully reduce the conservatism that typically accompanies CBF-based safety filters in MPC. The constructive character of the proofs and the resulting warm-start capability are practical strengths that could aid real-time deployment in safety-critical systems.

major comments (1)

[Abstract] The abstract asserts that the formulation is 'proven to improve feasibility and reachable sets with increasing prediction horizon' and that the proofs are constructive, yet the provided text supplies neither the statement of assumptions on the CBF or dynamics nor the key derivation steps establishing controlled invariance of the terminal set. This is load-bearing for the central claim.

minor comments (2)

[Simulations] The numerical validation reports a factor reduction in infeasible points but does not include error bars, multiple random seeds, or direct baseline comparisons against standard CBF-MPC without the terminal constraint; adding these would strengthen the empirical support.
[Notation and Preliminaries] Notation for the terminal CBF constraint and the warm-start procedure should be introduced with explicit definitions and cross-references to the relevant equations in the main text.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for their thorough review and valuable feedback on our manuscript. We have addressed the major comment point by point below, with revisions to enhance clarity and transparency regarding the central claims.

read point-by-point responses

Referee: [Abstract] The abstract asserts that the formulation is 'proven to improve feasibility and reachable sets with increasing prediction horizon' and that the proofs are constructive, yet the provided text supplies neither the statement of assumptions on the CBF or dynamics nor the key derivation steps establishing controlled invariance of the terminal set. This is load-bearing for the central claim.

Authors: We appreciate the referee identifying this gap in presentation. The manuscript's main body (Section II for system and CBF assumptions, Section IV for the constructive proof of terminal-set controlled invariance under the MPC policy, including the explicit derivation that the terminal CBF constraint remains invariant along closed-loop trajectories) does contain the required elements: the system is assumed control-affine with locally Lipschitz dynamics, the CBF is C^1 with relative degree one, and the terminal set is shown to be a controlled invariant set via a recursive feasibility argument that directly yields the horizon-dependent enlargement of the feasible region and reachable set. However, the abstract itself does not restate these assumptions or outline the key steps. We have therefore revised the abstract to include a concise statement of the assumptions and a high-level sketch of the invariance proof, while retaining the full derivations in the body. This change makes the central claim self-contained without altering any technical content. revision: yes

Circularity Check

0 steps flagged

No significant circularity detected

full rationale

The abstract describes an MPC formulation that imposes a given CBF as a terminal constraint, accompanied by constructive proofs that feasibility and reachable sets improve with longer horizons. No equations, fitted parameters, or self-referential definitions appear in the provided text. The proofs are explicitly characterized as constructive, indicating that the claimed improvements follow from the problem setup and invariance properties rather than reducing to a fit or a self-citation chain. No load-bearing step is shown to be equivalent to its own inputs by construction, and the numerical validation on a nonholonomic system is presented as external confirmation rather than a tautology. The derivation is therefore self-contained against external benchmarks.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Review based solely on abstract; no explicit free parameters, axioms, or invented entities are stated in the provided text.

pith-pipeline@v0.9.0 · 5486 in / 1039 out tokens · 95783 ms · 2026-05-08T07:31:38.321642+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

10 extracted references

[1]

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

Bansal, S., Chen, M., Herbert, S., and Tomlin, C. J., “Hamilton-jacobi reachability: A brief overview and recent advances,” 2017 IEEE 56th Annual Conference on Decision and Control (CDC), IEEE, 2017, pp. 2242–2253

2017
[2]

Discrete control barrier functions for safety-critical control of discrete systems with application to bipedal robot navigation

Agrawal, A., and Sreenath, K., “Discrete control barrier functions for safety-critical control of discrete systems with application to bipedal robot navigation.”Robotics: Science and Systems, Vol. 13, Cambridge, MA, USA, 2017, pp. 1–10

2017
[3]

Safety-critical model predictive control with discrete-time control barrier function,

Zeng, J., Zhang, B., and Sreenath, K., “Safety-critical model predictive control with discrete-time control barrier function,” 2021 American Control Conference (ACC), IEEE, 2021, pp. 3882–3889

2021
[4]

Enhancing feasibility and safety of nonlinear model predictive control with discrete-time control barrier functions,

Zeng, J., Li, Z., and Sreenath, K., “Enhancing feasibility and safety of nonlinear model predictive control with discrete-time control barrier functions,”2021 60th IEEE Conference on Decision and Control (CDC), IEEE, 2021, pp. 6137–6144

2021
[5]

Control barrier functions for cyber-physical systems and applications to NMPC,

Schilliger, J., Lew, T., Richards, S. M., Hänggi, S., Pavone, M., and Onder, C., “Control barrier functions for cyber-physical systems and applications to NMPC,”IEEE robotics and automation letters, Vol. 6, No. 4, 2021, pp. 8623–8630

2021
[6]

Predictive control barrier functions: Enhanced safety mechanisms for learning-based control,

Wabersich, K. P., and Zeilinger, M. N., “Predictive control barrier functions: Enhanced safety mechanisms for learning-based control,”IEEE Transactions on Automatic Control, Vol. 68, No. 5, 2022, pp. 2638–2651

2022
[7]

Discrete-time control barrier functions for guaranteed recursive feasibility in nonlinear mpc: An application to lane merging,

Katriniok, A., Shakhesi, E., and Heemels, W., “Discrete-time control barrier functions for guaranteed recursive feasibility in nonlinear mpc: An application to lane merging,”2023 62nd IEEE Conference on Decision and Control (CDC), IEEE, 2023, pp. 3776–3783

2023
[8]

Control barrier functions: Theory and applications,

Ames, A. D., Coogan, S., Egerstedt, M., Notomista, G., Sreenath, K., and Tabuada, P., “Control barrier functions: Theory and applications,”2019 18th European control conference (ECC), Ieee, 2019, pp. 3420–3431

2019
[9]

Feasibility enhancement of constrained receding horizon control using generalized control barrier function,

Ma, H., Zhang, X., Li, S. E., Lin, Z., Lyu, Y., and Zheng, S., “Feasibility enhancement of constrained receding horizon control using generalized control barrier function,”2021 4th IEEE International Conference on Industrial Cyber-Physical Systems (ICPS), IEEE, 2021, pp. 551–557

2021
[10]

CasADi: asoftwareframeworkfornonlinearoptimization and optimal control,

Andersson,J.A.,Gillis,J.,Horn,G.,Rawlings,J.B.,andDiehl,M.,“CasADi: asoftwareframeworkfornonlinearoptimization and optimal control,”Mathematical Programming Computation, Vol. 11, No. 1, 2019, pp. 1–36. 13

2019