( 05 )Layer

Human2Robot Retargeting

5 end-effectors

Human demonstrations converted into robot-executable trajectories, physics-replayed before delivery.

Wrist mapping, inverse kinematics and contact-preserving optimisation convert human demonstrations into robot-executable trajectories for Shadow, Inspire, Allegro, Franka/UR5 and Robotiq 3F.

Every retargeted trajectory is replayed in Isaac Sim or MuJoCo before delivery. Physically infeasible sequences are rejected, not shipped.

For other embodiments, send a URDF and we generate a custom retarget from the source MANO parameters — the five published end effectors are reference implementations, not a limit.

SPECIFICATION
End effectors
Shadow · Inspire · Allegro · Franka/UR5 · Robotiq 3F
Validation
Physics replay before delivery
Custom targets
From URDF, on request
Export
UMI · LeRobot · RDT · OXE

Available on MOVAS-Ego 100K. Coverage per scene family is published on each family page and on the data card.

FAQ

Human2Robot Retargeting — Common Questions

Send a URDF. We generate a custom retarget from the source MANO parameters. The five published end effectors are reference implementations rather than a constraint.

They are rejected and excluded from the delivery, and the rejection rate is reported on the data card. We do not ship sequences a robot cannot execute.

Your models need the physical world.

Start with training-ready data today. Scale to targeted re-capture when you’re ready.

Request a Subset

Data moves robots. MOVAS moves data.