Motion Planner Augmented Reinforcement Learning for Robot Manipulation in Obstructed Environments.
Jun YamadaYoungwoon LeeGautam SalhotraKarl PertschMax PfluegerGaurav S. SukhatmeJoseph J. LimPeter EnglertPublished in: CoRL (2020)
Keyphrases
- reinforcement learning
- humanoid robot
- motion control
- autonomous robots
- manipulation tasks
- mobile robot
- motion planning
- robot control
- robotic systems
- robotic tasks
- real robot
- robot motion
- inverse kinematics
- unstructured environments
- action selection mechanism
- dynamic environments
- autonomous navigation
- end effector
- position and orientation
- space time
- configuration space
- robot navigation
- human robot interaction
- robot behavior
- monocular vision
- image sequences
- action selection
- autonomous learning
- control signals
- parallel robot
- sagittal plane
- optical flow
- degrees of freedom
- hand eye calibration
- path planner
- motion estimation
- multi agent
- service robots
- motion model
- robot moves
- perceptual aliasing
- collision free
- learning algorithm
- real environment
- human motion
- robot arm
- path planning
- moving objects
- vision system
- robot teams
- camera motion
- physical constraints
- state space
- dynamic programming
- motion analysis
- function approximation
- model free
- machine learning
- optimal policy
- robot manipulators
- multi robot
- partially observable