analiza aiqbench
This research introduces ZeroTouch, a framework that enables robots to estimate contact forces and deformation using only visual data during deployment, eliminating the need for dedicated tactile sensors at the edge. By leveraging tactile signals solely as privileged supervision during training, the method significantly outperforms current large language model-based baselines like OpenVLA and SmolVLA in physical manipulation tasks. The work advances embodied AI by bridging the gap between visual observation and precise haptic feedback without requiring specialized hardware integration.
ZeroTouch: Tactile-Supervised Visual Contact Estimation for Contact-Rich Manipulation
arXiv:2609.21726v1 Announce Type: new Abstract: Reliable robotic grasping benefits from estimating the evolving physical interaction and selecting a grasp-dependent compression target. Tactile sensors provide direct interaction measurements but require dedicated hardware at deployment. We introduce ZeroTouch, a tactile-supervised framework that predicts dense contact deformation, the instantaneous six-axis wrench, and a grasp-dependent desired compression target from wrist RGB observations, gripper state, and local gravity direction. Tactile measurements are used only as privileged supervision during training and are not required at deployment. On the full validation set, the complete architecture reduces normal-force MAE from 2.017 N for a state-only baseline to 0.531 N. In physical evaluation with 20 trials per condition, ZeroTouch achieves 95% success on an unseen object, 80% in a seen-object/unseen-grasp condition, and 90% under a content/load shift. Under the same evaluation protocol, OpenVLA achieves 25%, 40%, and 55%, while SmolVLA achieves 10%, 25%, and 35%, respectively.
Abstract:Reliable robotic grasping benefits from estimating the evolving physical interaction and selecting a grasp-dependent compression target. Tactile sensors provide direct interaction measurements but require dedicated hardware at deployment. We introduce ZeroTouch, a tactile-supervised framework that predicts dense contact deformation, the instantaneous six-axis wrench, and a grasp-dependent desired compression target from wrist RGB observations, gripper state, and local gravity direction. Tactile measurements are used only as privileged supervision during training and are not required at deployment. On the full validation set, the complete architecture reduces normal-force MAE from 2.017 N for a state-only baseline to 0.531 N. In physical evaluation with 20 trials per condition, ZeroTouch achieves 95% success on an unseen object, 80% in a seen-object/unseen-grasp condition, and 90% under a content/load shift. Under the same evaluation protocol, OpenVLA achieves 25%, 40%, and 55%, while SmolVLA achieves 10%, 25%, and 35%, respectively.
Robotics (cs.RO)
arXiv:2609.21726 [cs.RO]
(or arXiv:2609.21726v1 [cs.RO] for this version)
https://doi.org/10.48550/arXiv.2609.21726 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Text integral sub licența deschisă a publicației · 235 cuvinte
Subiecte conexe
Tactile-JEPA: Topology-Aware Self-Supervised Representation Learning for Distributed Tactile Sensors
arXiv · 22 sept. 2026
TACO: TActile World Model as a Self-COrrector for Scalable Robot Policy Post-Training
arXiv · 17 sept. 2026
EgoTactile: Learning Grasp Pressure for Everyday Objects from Egocentric Video
arXiv · 9 iun. 2026
TacWAM: Anchor-Guided World Action Model with Mechanics-Aware Tactile Prediction
arXiv · 31 iul. 2026