Agility Robotics Trains Humanoid Robots with NVIDIA Simulation
Agility Robotics used NVIDIA Isaac Sim and Isaac Lab to train their Digit humanoid robot, reducing controller development from weeks to days through billions of simulation steps.
Impact
Weeks to days
Development cycle reduction
Billions of steps
Simulation training scale
Challenge
Developing reliable whole-body control for humanoid robots to operate safely alongside human workers in warehouse environments, with traditional development cycles taking weeks per iteration.
Solution
Adopted a simulation-first approach using NVIDIA Isaac Sim for digital twin modeling and Isaac Lab for reinforcement learning, running billions of parallel training episodes.
Tools & Technologies
What Leaders Say
“Isaac Sim running on NVIDIA GPUs lets us simulate years of real-world learning for Digit in just hours. That simulation speedup means we can train for all the conditions we might see on the factory floor.”
Sign up to read complete case studies, access detailed metrics, and unlock all use cases.
Full Story
U.S. warehouses face persistent labor gaps in material-handling operations. While humanoid robots offer a potential solution, they must operate with human-level agility in spaces not designed for machines. Achieving reliable whole-body control for safe operation alongside human workers presents inherent complexity.
Agility Robotics adopted a simulation-first approach using NVIDIA Isaac Sim to model Digit's chassis with OpenUSD, preserving joints, mass, and contact surfaces. Customer CAD and BIM data integrated into shared scene graphs with thousands of stress-test scenarios. Digital twin simulations fed into Isaac Lab, where millions of parallel reinforcement-learning episodes refined Digit's whole-body controller across billions of interactions.
The approach dramatically accelerated development: iteration cycles that previously took weeks were completed in days. Controllers trained in simulation transferred successfully to real warehouse and manufacturing operations with customers including GXO Logistics and Schaeffler.
Amazon Bedrock