MULTI-FINGERED HAND
Dexterity, power, and durability.
The only hand realized at human scale.
TECHNOLOGY
Technologies behind the multi-fingered hand
TECHNOLOGY / Details
Dexterity
16 DOF / 180 deg/s16 active joints / Finger-thumb opposability / Maximum joint velocity 180 deg/sec
Power
50N / 150sIt can exert a maximum fingertip force of 50 N and sustain it for up to 150 seconds.
Durability
450,000+ cyclesIt has completed 450,000 durability cycles under various fingertip load conditions, including 24,000 cycles involving lifting a 5 kg weight.
Impact Resistance
1.23m/sEven when the fingers hit an obstacle at 1.23m/s, the system can detect it and switch control mode to absorb the force.
Tactile sensor
288 channelsIn addition to six-axis force sensors at each fingertip, a total of 288 channels of tactile sensors are integrated into the fingers and palm, enabling high-precision contact detection and delicate manipulation.
FEATURES
While human hands are both dexterous and powerful, robots have struggled to achieve the same balance. Our multi-fingered hand provides the precision necessary to thread a needle, as well as an output of 50 N at a human scale. We have finally reached a practical level where it can “work like a human.”
WORK SCENES
Work Scenes
Publications
Published papers on multi-fingered hands
-
“R2-Dreamer: Redundancy-Reduced World Models without Decoders or Augmentation”,
In Proceedings of the International Conference on Learning Representations (ICLR), 2026.
-
“Relative Geometrical Constraint on Finger Motion for Dexterous Teleoperation of Multifingered Hand”,
IEEE/SICE International Symposium on System Integration (SII), 2026.
-
“Touch-Based Manipulation with Multi-Fingered Robot using Off-policy RL and Temporal Contrastive Learning”,
IEEE International Conferences on Robotics and Automation (ICRA), 2024.
-
“Quadratic Programming Based Inverse Kinematics for Precise Bimanual Manipulation”,
IEEE International Conferences on Robotics and Automation (ICRA), 2024.
-
“Offline Reinforcement Learning with Mixture of Deterministic Policies”,
Transactions on Machine Learning Research, 2023.
Overseas reseach branches
Research and development through a global network