Infusing model predictive control into meta-reinforcement learning for mobile robots in dynamic environments.
Jaeuk ShinAstghik HakobyanMingyu ParkYeoneung KimGihun KimInsoon YangPublished in: CoRR (2021)
Keyphrases
- dynamic environments
- mobile robot
- model predictive control
- reinforcement learning
- predictive control
- control system
- real robot
- path planning
- robot control
- reinforcement learning algorithms
- reinforcement learning agents
- obstacle avoidance
- autonomous agents
- mobile robotics
- robotic systems
- indoor environments
- collision avoidance
- autonomous navigation
- autonomous robots
- changing environment
- control scheme
- multi robot
- state space
- potential field
- multi agent
- motion planning
- unknown environments
- optimal policy
- learning algorithm
- highly dynamic environments
- agent systems
- single agent
- control strategy
- multiple models
- belief space
- control algorithm
- optimal control
- machine learning
- real time