Event
GR-VIP uses geometric representation conditioning to improve robot docking precision
1 reports · 1 sources
Chanhwi Lee and other researchers propose GR-VIP, a goal-conditioned visuomotor policy that uses a visual encoder pretrained for relative camera pose estimation to condition action prediction on dense relational features before pose regression, preserving fine-grained cross-view information to learn corrective actions from demonstrations. In real docking tasks, GR-VIP achieves high docking success rates and millimeter-level terminal precision in both the training environment and unseen environments, with transfer requiring no additional training. Supplementary experiments show that the policy can align with non-docking goals and complete docking under the low-light conditions tested. This work has been submitted to IEEE and is under review.
Report timeline
- GR-VIP: Geometric representation-conditioned visuomotor policy for precise docking
The study proposes GR-VIP, a goal-conditioned visuomotor policy that uses a visual encoder pretrained for relative camera pose estimation, conditioning action prediction on dense relational features before pose regression to preserve fine-grained cross-view information for learning corrective actions from demonstrations. In real-world docking tasks, GR-VIP achieves high docking success rates and millimeter-level terminal accuracy in both the training environment and unseen environments, with transfer requiring no additional training. Supplementary experiments show the policy can align with non-docking targets and complete docking under the low-light conditions tested.
arXiv cs.RO ·