PRM-RL: Long-range Robotic Navigation Tasks by Combining Reinforcement Learning and Sampling-based Planning.
Aleksandra FaustOscar RamirezMarek FiserKenneth OslundAnthony G. FrancisJames DavidsonLydia TapiaPublished in: CoRR (2017)
Keyphrases
- long range
- navigation tasks
- reinforcement learning
- motion planning
- mobile robot
- path planning
- short range
- real robot
- collision free
- belief space
- function approximation
- indoor environments
- dynamic environments
- robotic systems
- reinforcement learning algorithms
- conditional random fields
- multi robot
- action selection
- temporal difference
- model free
- rl algorithms
- collision avoidance
- long range correlations
- markov decision processes
- state space
- degrees of freedom
- long range interactions
- optimal policy
- multi agent
- dynamic programming
- partially observable
- machine learning
- learning algorithm
- markov random field
- computer vision
- partially observable domains
- action space
- higher order
- partially observable markov decision processes
- optimal path
- dynamical systems
- optimal control