TD-GRPC locomotion paper trains diverse Unitree H1-2 humanoid behaviors
Evidence notes
- TD-GRPC combines temporal-difference learning with a group relative policy constraint for humanoid locomotion.
- The paper shows Unitree H1-2 tasks including standing, sitting, walking, running, crawling, stair climbing, balancing, and pole navigation.
Company context
Develops quadruped and humanoid robots, with strength in dynamic locomotion, vertically integrated hardware, and relatively low-cost commercial deployment. Unitree is one of the clearest examples of a legged robotics company moving from research visibility into real productisation and broader market distribution.