ContactGuard: Pre-Contact Execution Monitoring with Action-Conditioned Latent World Models
Contact-rich manipulation failures are often detected only after the robot has committed to contact. This is especially limiting in wrist-camera setups: close gripper--object views help observe contact, but a poor approach may already push, miss, slip, or disturb the object before conventional detectors react. We introduce \emph{ContactGuard}, a pre-contact execution monitor for chunked visuomotor policies. Given the policy's planned action chunk, ContactGuard predicts its short-horizon consequence in latent visual space and aborts if the predicted future latent indicates likely failure. Its l
Lineage graph
Paper → model → repo connections mined from source citations (Tier-1 exact match).
Why these links exist
Every edge carries a method, confidence, and the source snippet that justified it — so bad links are debuggable.
- PossiblePossibly related (embedding) · 48%Why first person video may matter for robot learning[D] →
- PossiblePossibly related (embedding) · 46%Video Friday: AI Gives Robot Hands Humanlike Dexterity →
- PossiblePossibly related (embedding) · 46%Video Friday: Robotic Motion Discovery Reveals Unusual Behaviors →
- FuzzySimilar title/name (fuzzy) · 84%liguodongiot/llm-action →
“Fuzzy title match (0.92): “ContactGuard: Pre-Contact Execution Monitoring with Action-C” ≈ “liguodongiot/llm-action””
- LinkedLinked via arxiv author · 85%Gehan Zheng →
“ContactGuard: Pre-Contact Execution Monitoring with Action-Conditioned Latent World Models”
- LinkedLinked via arxiv author · 85%Matthew Johnson-Roberson →
“ContactGuard: Pre-Contact Execution Monitoring with Action-Conditioned Latent World Models”
- LinkedLinked via arxiv author · 85%Weiming Zhi →
“ContactGuard: Pre-Contact Execution Monitoring with Action-Conditioned Latent World Models”