Research
* equal contribution ✉ corresponding author
Preprints
★ Taccel: Scaling Up Vision-based Tactile Robotics via High-performance GPU Simulation
Yuyang Li * * * Puhao Li Zihang Zhao Tengyu Liu ✉ Yixin Zhu ✉ Siyuan Huang ✉
Taccel is a high-performance GPU-based simulator, combining ABD and IPC, for simulating robots with vision-based tactile sensors.
Paper Code Website / Docs VideoPreprint
Afford-X: A Generalizable and Slim Affordance Reasoning Framework for Task-oriented Manipulation
* Yuyang Li * Huan-ang Gao Yixin Zhu ✉ Hao Zhao ✉
We develop Afford-X, an end-to-end trainable affordance reasoning model that incorporates Verb Attention and Bi-Fusion modules to improve multi-modal understanding. We showcase Afford-X's effectiveness in enabling task-oriented manipulations for robots across various tasks and environments, underscoring its efficiency and broad implications for advancing robotics and AI systems in real-world applications.
arXiv Code Video WebsitePreprint
Publications
★ Embedding high-resolution touch across robotic hands enables adaptive human-like grasping
Zihang Zhao * Wanlin Li * Yuyang Li * Tengyu Liu * Boren Li Meng Wang Hangxin Liu ✉ Yixin Zhu ✉ Qining Wang Kaspar Althoefer ✉ Song-Chun Zhu
We design F-TAC Hand, a five-finger biomimetic hand featuring high-resolution tactile sensing across the hand. This design enables adaptive human-like grasping and unlocks more capabilities.
arXiv Code Website Video 人民日报 新华网 中国新闻网 光明网 TechXplore 中国科技网 中国科学报 北大新闻网 北大新工科 北大AI院官微 北大AI院官网Nature Machine Intelligence
RoboVerse
Haoran Geng * * * Yuyang Li * * * * Yiran Geng Hao Dong Siyuan Huang
A comprehensive framework comprising a simulation platform, a synthetic dataset, and unified benchmarks.
Paper Code Website DocsRobotics: Science and Systems (R:SS) 2025
★ Tac-Man: Tactile-Informed Prior-Free Manipulation of Articulated Objects
Zihang Zhao * Yuyang Li * Wanlin Li Zhenghao Qi Lecheng Ruan ✉ Yixin Zhu ✉ Kaspar Althoefer
A prior-free method for articulated object manipulation, based on tactile pattern from GelSight-inspired sensors.
Website arXiv PDF Code VideoIEEE Transactions on Robotics (T-RO)
ManipTrans: Efficient Dexterous Bimanual Manipulation Transfer via Residual Learning
Puhao Li Tengyu Liu Yuyang Li Siyuan Huang ✉
A novel two-stage method for efficiently transferring human bimanual skills to dexterous robotic hands in simulation.
Website arXiv Code DataCVPR 2025
Ag2Manip: Learning Novel Manipulation Skills with Agent-Agnostic Visual and Action Representations
Puhao Li * Tengyu Liu * Yuyang Li Muzhi Han Haoran Geng Shu Wang Yixin Zhu Song-Chun Zhu Siyuan Huang ✉
Learning novel manipulation skills with agent-agnostic value and action representations.
Website arXiv PDF Code Video PosterIROS 2024, oral pitch presentation
PreAfford: Affordance-based Pre-grasping across Objects and Scenes
* Yuyang Li Huan-ang Gao Yixin Zhu Guyue Zhou Hao Dong Hao Zhao ✉
A universal pre-grasping manipulation framework using affordance representation.
Website arXiv PDF Video PosterIROS 2024, oral pitch presentation
★ Grasp Multiple Objects with One Hand
Yuyang Li Bo Liu Yiran Geng Puhao Li Yaodong Yang Yixin Zhu Tengyu Liu ✉ Siyuan Huang ✉
We propose MultiGrasp, a two-stage framework for simultaneous multi-object grasping with multi-finger dexterous hands. In addition, we contribute Grasp'Em, a large-scale synthetic multi-object grasping dataset.
Website arXiv PDF Code Data VideoIEEE Robotics and Automation Letter (RA-L) | IROS 2024, oral presentation
GenDexGrasp: Generalizable Dexterous Grasping
Puhao Li * Tengyu Liu * Yuyang Li Yiran Geng Yaodong Yang Yixin Zhu Siyuan Huang ✉
We propose GenDexGrasp, a versatile dexterous grasping method that can generalize to out-of-domain robotic hands. In addition, we contribute MultiDex, a large-scale synthetic dexterous grasping dataset.
Website arXiv PDF CodeICRA 2023