R-SARL: Crowd-aware Navigation Based Deep Reinforcement Learning for Nonholonomic Robot in Complex Environments.
Yanying ZhouShijie LiJochen GarckePublished in: CoRR (2021)
Keyphrases
- complex environments
- mobile robot
- obstacle avoidance
- reinforcement learning
- path planning
- indoor environments
- trajectory planning
- motion planning
- real robot
- learning agents
- autonomous navigation
- autonomous vehicles
- collision free
- agent learns
- navigation tasks
- unknown environments
- collision avoidance
- robot control
- visual navigation
- sensory information
- robot navigation
- outdoor environments
- sequential decision tasks
- autonomous robots
- autonomous agents
- dynamic environments
- function approximation
- service robots
- rough terrain
- robotic systems
- human robot interaction
- multi robot
- degrees of freedom
- multi agent
- humanoid robot
- topological map
- state space
- robot manipulators
- perceptual aliasing
- reinforcement learning algorithms
- robotic arm
- autonomous control
- cognitive systems
- model free
- dead reckoning
- dynamic programming
- temporal difference
- crowd sourced
- learning algorithm
- action selection
- control law
- sensory data