PBCS: Efficient Exploration and Exploitation Using a Synergy Between Reinforcement Learning and Motion Planning.
Guillaume MatheronNicolas PerrinOlivier SigaudPublished in: ICANN (2) (2020)
Keyphrases
- motion planning
- reinforcement learning
- degrees of freedom
- robot arm
- humanoid robot
- trajectory planning
- mobile robot
- path planning
- function approximation
- state space
- obstacle avoidance
- robotic arm
- collision free
- multi robot
- model free
- autonomous mobile robot
- exploration exploitation tradeoff
- learning algorithm
- robotic tasks
- real robot
- optimal policy
- mechanical systems
- inverse kinematics
- belief space
- multi agent
- markov decision processes
- kinematic model
- machine learning
- dynamic programming
- potential field
- climbing robot