Evidence item

The passage directly reports the two-scene UR10 setup, 3-cm convergence, collision-free paths, and image-only collision-free comparison rates. DATASET_RESULT

Authenticity UNVERIFIED · extraction UNVERIFIED · independence group none (unreviewed) · created seq 16198

Source: SE(3) Neural Potential Fields for 6-DoF Trajectory Planning Directly from Images Without Explicit 3D Reconstruction · locator {} · excerpt hash sha256:dd5a21f0ad3d3039ed4450231fee726a2639f3f752d3dc2e904d9b31b1203b9d

Exact excerpt

We present an SE(3) neural potential field learned from posed RGB images and supervised with a navigation function, the geodesic distance to the grasp through free space recovered from those same images during training, which removes both failures. On two tabletop scenes, from obstacle-blocked starts executed on a UR10, the field converges within 3 cm of the grasp from every start and every path it executes is collision-free against the ground-truth geometry, against 25% and 0% under image supervision alone; mean clearance rises from under a centimeter to 8.6-8.8 cm and arm-link contacts fall from 20.6-50.4% to 2.7-5.5% of executed configurations.

Linked claims

Claim	Direction	Relevance	Steps	Counted
00a3f605	SUPPORT	DIRECT	0	yes

Audits

None.