RMRL: Robot Navigation in Crowd Environments With Risk Map-Based Deep Reinforcement Learning.
Haodong YangChenpeng YaoChengju LiuQijun ChenPublished in: IEEE Robotics Autom. Lett. (2023)
Keyphrases
- robot navigation
- autonomous robots
- map building
- reinforcement learning
- continuous state
- topological map
- robot control
- autonomous mobile robot
- robot localization
- mobile robot
- robotic systems
- real time stereo
- scene understanding
- risk management
- function approximation
- dynamic environments
- model free
- state space
- optimal policy
- initially unknown
- activity recognition
- em algorithm
- image sequences
- learning algorithm