Paper page - CARVE: Certified Affordable Repair of Vetoed Maneuvers via Envelopes for Interactive Driving

Source: https://huggingface.co/papers/2606.02641

Abstract

Interactive driving scenarios reveal failures in autonomous vehicle rule systems where small adjustments could restore feasibility, leading to the introduction of CARVE, a certification framework that provides runtime proofs for multi-agent repairs without requiring predictions of other drivers’ compliance.

Interactive drivingexposes a failure mode that is easy to miss in rule-aware autonomous-driving stacks: a hard-rule margin can be negative for an ego candidate even though a small lawful accommodation by a non-priority agent would restore feasibility. Existingrulebooks,shields, andreachability filters are strong at vetoing unsafe actions, whileprediction-based planners model likely responses. Neither returns a runtime proof object that states whichbounded multi-agent edit repairsthe maneuver, who owns the edit, whether the request is right-of-way affordable, and what ego fallback remains if the request is not observed. We formulate this missing object as *interactive repair certification* and introduce *CARVE*, a prediction-free certificate layer over afinite latticeof ego-owned andagent-owned tactical operators. Agent-owned requests are admissible only inside \(B_j(s) = β(π_j)α_j^{\max}(s)\), acooperation envelopethat separates kinematic reachability from normative priority. The resulting certificate records the binding rule, repair category, repair set, responsibility-weighted cost split, and fallback. On 589 Lanelet2-geometry-grounded INTERACTION replay episodes,CARVE-Greedy accepts 98.64% of initially vetoed maneuvers and recovers 370/378 human-resolved false vetoes, while preserving 589/589right-of-way respect, zero priority-agent false positives, and 400/400 negative-stress vetoes. We provecertificate soundness, structuralright-of-way respect, exact finite-lattice minimality,fallback contingency, andblame-consistencyconditions.CARVEdoes not predict or require another driver’s compliance; it certifies whether a proposed interaction is bounded, attributable, and normatively admissible under declared assumptions.

View arXiv pageView PDFProject pageAdd to collection

Get this paper in your agent:

hf papers read 2606\.02641

Don’t have the latest CLI?curl \-LsSf https://hf\.co/cli/install\.sh \| bash

Models citing this paper0

No model linking this paper

Cite arxiv.org/abs/2606.02641 in a model README.md to link it from this page.

Datasets citing this paper0

No dataset linking this paper

Cite arxiv.org/abs/2606.02641 in a dataset README.md to link it from this page.

Spaces citing this paper0

No Space linking this paper

Cite arxiv.org/abs/2606.02641 in a Space README.md to link it from this page.

Collections including this paper0

No Collection including this paper

Add this paper to acollectionto link it from this page.