Back to search
2109.09047

Model-Free Safety-Critical Control for Robotic Systems

Tamas G. Molnar, Ryan K. Cosner, Andrew W. Singletary, Wyatt Ubellacker, Aaron D. Ames

correcthigh confidence
Category
Not specified
Journal tier
Strong Field
Processed
Sep 28, 2025, 12:56 AM

Audit review

The paper’s Theorem 3 constructs the same Lyapunov–barrier function hV(q,ė) = −V(q,ė) + αe h(q) with αe = ((λ−α)k1)/Ch and shows ḣV ≥ −α hV on SV, yielding forward invariance by a CBF-style argument (via Theorem 2) . The candidate solution uses the identical hV and constants, proves ḣV ≥ 0 on the boundary hV=0, and applies a first-exit-time contradiction to establish invariance—an equivalent argument. Minor differences are present in how invariance is certified (global differential inequality vs. boundary condition), but the logic, assumptions, and result coincide with the paper’s proof. The only omissions in the model’s write-up are small: it does not restate the paper’s nonzero-gradient-on-S condition for h and the standard existence/uniqueness assumptions used in the paper’s preliminaries .

Referee report (LaTeX)

\textbf{Recommendation:} minor revisions

\textbf{Journal Tier:} strong field

\textbf{Justification:}

The theorem and its proof are technically sound and provide a practical, general mechanism to ensure safety by tracking a kinematically safe velocity. The contribution is meaningful in robotics where full dynamic models are costly. Minor clarifications on assumptions and invariance mechanics would make the presentation even clearer to broad audiences.