Reading room
Annotated research paper shelf covering tactile gloves, wearable sensing rigs, datasets, and human-to-robot dexterous transfer.
Suting Ni et al. · ShanghaiTech University
A benchmark that pushes human-to-robot dexterous transfer past kinematic mimicry: whole-hand tactile signals aligned with hand-object interaction data.
Jiaming Liu et al. · Peking University
Instead of retargeting human demos geometrically, LaST-HD trains an action-conditioned world model on unpaired human and robot trajectories.
Changyi Lin, Ding Zhao · Carnegie Mellon University
A glove built from 16 rigid functional surfaces capturing 22-DoF surface motion and 2048 taxels of tactile contact.
Haoyang Luo et al.
Eighteen IMUs and mechanical stress isolation give full-hand tracking at 800 Hz to capture real motion energy above 100 Hz.
Hao-Shu Fang et al. · MIT & UC Berkeley
A passive perioperation exoskeleton with whole-hand tactile sensing for contact-rich, long-horizon bimanual demonstrations.
Jiaxin Lu et al. · UT Austin & Adobe Research
Mocap human-object interaction sequences with detailed hands manipulating objects paired with natural-language scripts.
Han Zhang et al. · Tsinghua University
Open-source glove with 21-DoF motion capture and 5-DoF haptic force feedback for blind dexterous teleoperation.
Chen Wang et al. · Stanford University
Portable hand-mocap rig paired with DexIL to retarget wild human manipulation data to robot embodiments.
Joseph DelPreto et al. · MIT CSAIL
Full wearable sensor suite capturing eye tracking, forearm EMG, 17-IMU tracking, and tactile gloves for daily kitchen tasks.
Oliver Glauser et al. · ETH Zurich & NYU
Capacitive stretch-sensor array soft glove with deep neural network reconstruction of hand pose.