Event
Robot-to-Human Bimanual Handover: Combining Tactile and Compliant Control
1 reports · 1 sources
Pasquale Marra and other authors propose a robot-to-human bimanual handover method that combines a vision-language-action (VLA) model with a compliant controller, fine-tuning the model using RGB observations, temporally encoded tactile feedback, and proprioceptive human demonstrations; the compliant controller is used to reduce interaction forces during object handover. In a human-subject study, the authors compared the full system with two baselines: one without tactile feedback, and another with tactile feedback but without compliant control. Results show that compliant control and tactile history are complementary, and the full system clearly outperforms both baselines. The project page has been made public.
Report timeline
- Temporal Tactile Encoding and Compliant Control for Intent-Aware Robot-to-Human Bimanual Handover
This study combines a VLA model with a compliant controller, fine-tuning the model using RGB observations, temporally encoded tactile feedback, and proprioception from human demonstrations to achieve robot-to-human bimanual handover. In a human-subject study, the system was compared against two baselines: one without tactile feedback and one with tactile feedback but no compliant control. Results show that compliant control and tactile history are complementary, with performance clearly outperforming the baselines. The project page is publicly available.
arXiv cs.RO ·