FeasibleCap: Real-Time Embodiment Constraint Guidance for In-the-Wild Robot Demonstration Collection

Dimensions overall score 8.0

GitHub Code Pulse

No public code linked for this paper yet.

• FeasibleCap provides real-time executability feedback during gripper-in-hand data collection without requiring head-worn AR/VR displays.method 95%
• FeasibleCap improves replay success rates on pick-and-place and tossing tasks compared to baseline collection methods.result 90%
• FeasibleCap reduces the fraction of infeasible frames during demonstration collection, with the largest gains on tossing tasks.result 90%
• FeasibleCap checks reachability, joint-rate limits, and collisions against a target robot model at each frame during collection.technical 95%
• FeasibleCap provides feedback through on-device visual overlays and haptic cues without requiring learned models or robot hardware.method 95%
• Enforcing executability constraints during collection does not sacrifice cross-embodiment transfer across robot platforms.result 85%
• Existing collection-time feedback systems rely on head-worn AR/VR displays, robot-in-the-loop hardware, or learned dynamics models.limitation 90%
• Failed demonstrations inflate the effective cost per usable trajectory through repeated collection, diagnosis, and validation.technical 90%

Competitive landscape

Competitor map is still being generated for this paper. Enable generation or check back soon.

Startup potential card