Boston Dynamics and RAI Institute Partner to Advance Humanoid Robots Through Reinforcement Learning
The partnership focuses on using reinforcement learning to build dynamic and generalizable capabilities on the Atlas robot. Objectives include developing sim-to-real mobility, improving whole-body loco-manipulation, and exploring full-body contact strategies. This collaboration builds upon previous joint development work on Spot’s Reinforcement Learning Researcher Kit.
Evidence notes
- The partnership focuses on using reinforcement learning to build dynamic and generalizable capabilities on the Atlas robot.
- Objectives include developing sim-to-real mobility, improving whole-body loco-manipulation, and exploring full-body contact strategies.
- This collaboration builds upon previous joint development work on Spot’s Reinforcement Learning Researcher Kit.
Company context
Boston Dynamics builds high-performance mobile robots including Spot for industrial and field inspection, Stretch for warehouse trailer unloading, and Atlas for advanced humanoid mobility.