Event
Crab robot achieves dynamic jumping via reference-guided residual learning
1 reports · 1 sources
The Jumper project uses a reference-guided residual reinforcement learning method to control a crab robot in performing dynamic jumps. Its approach is: the reference motion records only the nominal trajectory, and the policy learns only the correction to the reference commands, rather than generating motions from scratch. According to the description of this method, the push-off impulse determines the takeoff velocity, and the ballistic trajectory determines the airborne phase; the project completed jumping and boundary tests with the official policy.
Report timeline
- Jumper: Dynamic Jumping Control for a Crab Robot Based on Reference Motion and Residual Reinforcement Learning
Jumper uses a reference-guided residual reinforcement learning method to control dynamic jumping in a crab robot. The reference motion records only the nominal trajectory, and the policy learns only the corrections to the reference commands. In this method, the takeoff impulse determines the launch velocity, the ballistic trajectory determines the airborne phase, and the official policy is used to complete jumping and boundary tests.
Guyuehome ·