VLM-Safe-RL adds frozen VLM signals as anticipatory costs to the CMDP Lagrangian update via dual-path CLIP, VLM-Lagrange, and confidence gating, outperforming baselines on Safety-Gymnasium FormulaOne while showing partial generalization.
Title resolution pending
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
years
2026 2verdicts
UNVERDICTED 2representative citing papers
A monocular SLAM-plus-semantics pipeline builds a class-inflated ESDF that supplies distance and gradient inputs to a CBF controller for online safe teleoperation and navigation.
citing papers explorer
-
Seeing Before Colliding: Anticipatory Safe RL with Frozen Vision-Language Models
VLM-Safe-RL adds frozen VLM signals as anticipatory costs to the CMDP Lagrangian update via dual-path CLIP, VLM-Lagrange, and confidence gating, outperforming baselines on Safety-Gymnasium FormulaOne while showing partial generalization.
-
Embedding Semantic Risk into Distance Fields and CBFs for Online Monocular Safe Control
A monocular SLAM-plus-semantics pipeline builds a class-inflated ESDF that supplies distance and gradient inputs to a CBF controller for online safe teleoperation and navigation.