A reinforcement learning curriculum with self-generated prior policies enables a simulated quadruped to hop on one leg over gaps up to 60 cm and stepping stones spaced 15 to 35 cm apart.
Transferable latent-to-latent locomotion policy for efficient and versatile motion control of diverse legged robots,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.RO 1years
2025 1verdicts
CONDITIONAL 1roles
background 1polarities
support 1representative citing papers
citing papers explorer
-
Jump-Start Reinforcement Learning with Self-Evolving Priors for Extreme Monopedal Locomotion
A reinforcement learning curriculum with self-generated prior policies enables a simulated quadruped to hop on one leg over gaps up to 60 cm and stepping stones spaced 15 to 35 cm apart.