CARL: controllable agent with reinforcement learning for quadruped locomotion.
Ying-Sheng LuoJonathan Hans SoesenoTrista Pei-Chun ChenWei-Chao ChenPublished in: ACM Trans. Graph. (2020)
Keyphrases
- reinforcement learning
- multi agent
- action selection
- quadruped robot
- legged robots
- multi agent systems
- state abstraction
- learning agent
- function approximation
- partially observable
- state action
- reinforcement learning agents
- markov decision process
- learning capabilities
- multiple agents
- autonomous agents
- autonomous learning
- robot control
- intelligent agents
- agent receives
- software agents
- reward function
- decision making
- learning agents
- mobile agents
- single agent
- state space
- multi agent environments
- exploration strategy
- cooperative
- agent model
- agent architecture
- agent learns
- degrees of freedom
- multiagent systems
- reward shaping
- rough terrain
- machine learning
- reinforcement learning algorithms
- learning algorithm
- agent technology
- mobile robot
- dynamic programming
- multi agent learning
- action space
- multi agent reinforcement learning
- real robot
- model free
- optimal control
- robotic systems
- learning process