Robot navigation in crowds via deep reinforcement learning with modeling of obstacle uni-action.
Xiaojun LuHanwool WooAngela FaragassoAtsushi YamashitaHajime AsamaPublished in: Adv. Robotics (2023)
Keyphrases
- robot navigation
- reinforcement learning
- continuous state
- autonomous robots
- autonomous mobile robot
- mobile robot
- scene understanding
- action selection
- real time stereo
- map building
- initially unknown
- partially observable domains
- feature selection
- state action
- action space
- markov decision processes
- markov decision process
- model free
- optimal policy
- optical flow
- landmark recognition
- image sequences