arXiv:2512. 10946v2 Announce Type: replace-cross Abstract: Human-level contact-rich manipulation relies on the distinct roles of two key modalities: vision provides spatially rich but temporally slow global context, while force sensing captures rapid local contact dynamics.
By Wendi Chen, Han Xue, Yi Wang, Fangyuan Zhou, Jun Lv, Yang Jin, Shirun Tang, Chuan Wen, Cewu Lu
arXiv:2609.01596v1 Announce Type: cross
Abstract: Real-world robotic assembly at sub-millimeter tolerances demands spatial precision, compliant interaction, and robustness to contact failures. We pre...
By Haoyuan Deng, Haichao Liu, Wenkai Guo, Yuan Ling, Zaijia Yang, Yuanjiang Xue, Haosheng Sun, Liangzi Wang, Ziwei Wang
The paper introduces an interaction‑centric framework that unifies representations for two‑finger gripper manipulation across different robot embodiments. By using a parameterized universal gripper abstraction and a canonical gripper‑frame representation, the system infers sub‑tasks from language and RGB‑D inputs, grounds interaction triplets, and employs hybrid features and a Flow‑Matching Transformer to generate smooth 7‑DoF action sequences. Experiments in both simulation and real‑world settings show that this approach achieves competitive benchmark performance while enabling extreme cross‑embodiment and cross‑viewpoint zero‑shot sim‑to‑real transfer to heterogeneous robot platforms.
By Guanlin Li, Shifeng Bao, Yihan Zhao, Haitao Shen, Haoyang Li, Chen Zhao, Tong Yang, Jie Tang, Jing Zhang
Real-world robotic assembly at sub-millimeter tolerances demands spatial precision, compliant interaction, and robustness to contact failures. We present Facet-0, a robotic foundation model that predi...
arXiv:2511. 14427v4 Announce Type: replace-cross Abstract: Effective contact-rich manipulation requires robots to synergistically leverage vision, force, and proprioception.
By Rickmer Krohn, Vignesh Prasad, Gabriele Tiboni, Georgia Chalvatzaki
The paper presents a method for force‑aware manipulation that relies solely on visual predictions of force from a compliant Fin Ray gripper. It trains a visual force estimator on calibration data to annotate demonstrations, then uses these annotated demonstrations to train an action–force proposal policy that generates candidate actions and their expected forces. At deployment, the robot selects the action whose predicted force best matches a target force from the demonstrations, enabling contact‑rich tasks such as berry picking, can grasping, in‑hand reorientation, and plug insertion without physical force or tactile sensors.
By Haonan Chen, Feiyang Wu, Yuxiang Ma, Mustafa Mete, Pengfei Ye, Junxuan Shen, Cheng Zhu, Aurora Ruggeri, Kelvin Cheung, Jiayuan Mao, Edward Adelson, Jiajun Wu, Robert D. Howe, Yilun Du