Occlusion-Safe Shared Micromanipulation in Vision-Constrained Environments
Résumé
Vision-constrained teleoperation systems are vulnerable to occlusions, which can affect closed-loop control and compromise both performance and safety. We present an occlusion-safe shared-control framework that prevents robots from entering unobserved regions under fixed-camera monitoring. Our method combines Control Lyapunov Functions (CLFs) and High-Order Control Barrier Functions (HOCBFs) within a Quadratic Programming (QP) solver to enforce task goals and occlusion-avoidance constraints in real time. A human operator provides high-level commands through a 6-DoF haptic interface, while an autonomous safety layer ensures continuous visibility. To help the operator understand the system behavior, we deliver haptic feedback that signals proximity to occlusions and provide visual overlays supported by a virtual reality camera for improved scene understanding.
Domaines
| Origine | Fichiers produits par l'(les) auteur(s) |
|---|---|
| Licence |