Advanced Telerobotics Research Lab
Department of Computer Science, Kent State University
ATR Lab logo Kent State Computer Science logo
Hands-on workshop

Teach a robot your dance

Motion retargeting and robot learning. A phone video of someone dancing becomes 3D human motion, gets retargeted to the Booster K1 humanoid, trained with reinforcement learning in simulation, and runs on the real robot.

Date
Saturday, October 24, 2026
Time
10:00 AM – 2:00 PM
Place
MSB 228, Kent State
The real Booster K1 humanoid robot jumping in the lab Real robot

How it works

From video to robot
3D human body model with blue keypoints
1. 3D bodyMotion capture
K1 robot in simulation doing the same jump
2. Robot in simulationRetargeted motion
Trained policy in simulation with blue reference points
3. Trained policyRL in simulation
  1. 01

    Motion capture

    Your video becomes a 3D body moving in space.

  2. 02

    Retargeting

    The human motion is mapped onto the robot's joints and proportions.

  3. 03

    Training in simulation

    Reinforcement learning teaches the robot to follow the motion without falling.

  4. 04

    Sim-to-real

    The trained policy runs on the real K1.

The real K1 robot dancing The real K1 robot dancing The real K1 robot dancing

More dances on the real K1.

Tools: GVHMR (video to 3D motion), GMR (retargeting), MuJoCo and mjlab (simulation and training).

What to bring

  • A laptop.
  • A short video of yourself dancing is nice to have, but optional.