章节结构（全文标题提取）： 1 Introduction 2 Related Work 3 Latent World Models for Pre-Contact Grasp Monitoring 4 Experimental Results 5 Conclusions, Limitations, and Future Work 3.1 Problem Setting and Data · 3.2 Background: LeWM-style Latent Prediction · 3.3 Multi-View Latent World Model · 3.4 Linear Failure Probe on Predicted Latents · 3.5 Online Pre-Contact Grasp Monitor · 4.1 Experimental Setup · 4.2 Closed-Loop Grasp-Outcome Prediction · 4.3 Offline Diagnostics: What Information Does the Monitor Use?
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 latent world model is trained from unlabelled robot trajectories to predict compact multi-view visual embeddings under planned actions, avoiding pixel-level video prediction. A lightweight failure probe is then trained from a small labelled set of pre-contact clips. At deployment, ContactGuard anchors prediction before an imminent contact event, rolls the model forward under the policy's own actions, and verifies the predicted post-contact latent. Across real-world contact-rich manipulation tasks, ContactGuard predicts failure more accurately than direct and corrupted-action ablations, and transfers to live robot as a pre-contact abort signal without modifying the underlying policy.