Physics-Based Motion Control Through DRL's Reward Functions.
Antonio Santos de SousaRubens Fernandes NunesCreto Augusto VidalJoaquim Bento Cavalcante NetoDanilo Borges da SilvaPublished in: SVR (2021)
Keyphrases
- motion control
- reward function
- mobile robot
- markov decision processes
- reinforcement learning
- control system
- state space
- inverse reinforcement learning
- control policies
- physical constraints
- optimal policy
- robot control
- autonomous robots
- transition probabilities
- multiple agents
- state variables
- policy search
- markov decision process
- autonomous navigation
- path planning
- generative model
- kinematic model
- real robot
- function approximation
- markov chain