DEXTEROUS MANIPULATION
Compare hardware. Discover applications. Share what works. An open reference and community for dexterous manipulation.
Active / total DOF: 20 / 24
Joint position; 40 tendon-load sensors; 1 IMU; 2 standard / up to 5 Shadow Tactile Fingertips (STF)
Curated public references, checked 2026-09-22. Links describe a model or family relationship, not hardware-tested compatibility.
DexSuite · Family resource
Research-maintained visual and collision models for multiple hand families.
Inspect the model directory and upstream asset license; revision, size and sensor geometry may differ from your hand.
DexSuite · Family resource
Human-to-robot motion optimizers with configurations for the listed hand families.
Retargeting generates joint targets; it is not a device driver. Verify model geometry and joint order.
Shadow Robot · Family resource
Manufacturer control interfaces and integration examples.
Select the matching hand configuration and ROS branch; Lite variants have different joints.
Shadow Robot · Family resource
URDF descriptions and common packages for Shadow hardware.
Check joint count, hand side and hardware revision before using a model.
Google DeepMind · Model-specific
Simulation assets for the E3M5 Shadow Hand, with left and right versions.
A simulation model, not a hardware driver. Match the E3M5 geometry to your unit.
PKU EPIC · Research use
Synthetic grasp data and generation code using a Shadow Hand model.
Simulation-generated grasps are not validated hardware executions.
PKU EPIC · Research use
Vision-based dexterous grasp learning with Shadow Hand simulation tasks.
Follow the authors’ environment and model versions; deployment requires additional work.
DFKI Robotics Innovation Center · Research use
Research code for transferring human motions to several supported robot hands.
Use the project’s specific model, calibration and motion-capture assumptions; mesh licenses vary.
