Actuator-Constrained Reinforcement Learning for High-Speed Quadrupedal Locomotion.
Young-Ha ShinTae-Gyu SongGwanghyeon JiHae-Won ParkPublished in: CoRR (2023)
Keyphrases
- high speed
- reinforcement learning
- legged robots
- low power
- robot control
- function approximation
- state space
- markov decision processes
- frame rate
- high speed networks
- learning algorithm
- mobile robot
- degrees of freedom
- machine learning
- real robot
- reinforcement learning algorithms
- policy search
- multi agent
- real time
- control system
- model free
- learning process
- stochastic approximation
- robotic control
- temporal difference learning
- closed loop
- pose estimation
- optimal policy
- low cost