Motion Planning as Online Learning: A Multi-Armed Bandit Approach to Kinodynamic Sampling-Based Planning.
Marco FaroniDmitry BerensonPublished in: IEEE Robotics Autom. Lett. (2023)
Keyphrases
- motion planning
- online learning
- multi armed bandit
- degrees of freedom
- mobile robot
- path planning
- trajectory planning
- regret bounds
- robotic tasks
- multi armed bandits
- belief space
- humanoid robot
- multi robot
- robotic arm
- collision free
- e learning
- mechanical systems
- active learning
- reinforcement learning
- climbing robot
- decentralized decision making
- linear regression
- multi class
- decision trees