WPP Slashes Humanoid Robot Training Time by 10x Using NVIDIA and Google Cloud Technology
Summary
WPP slashes humanoid robot training time from 24 hours to under one hour using NVIDIA RTX PRO 6000 Blackwell GPUs on Google Cloud, while a cutting-edge motion capture and digital twin workflow enables robots to mimic natural human movement with remarkable stability.
Key Points
- WPP uses Google Cloud's G4 VMs powered by NVIDIA RTX PRO 6000 Blackwell to train humanoid robots, cutting reinforcement learning training cycles from 24 hours down to under one hour — a 10x speed improvement.
- A motion capture and digital twin workflow using OptiTrack, OpenUSD, and MuJoCo physics simulation is used to remap complex human movement onto robots with far fewer degrees of freedom, enabling natural and stable physical motion.
- Unitree has released its reinforcement learning code on GitHub alongside the NVIDIA Isaac Sim image on Google Cloud Marketplace, making advanced humanoid robot motion research widely accessible to developers.