Coarse planning for landmark navigation in a neural-network reinforcement-learning robot.
Gianluca BaldassarrePublished in: IROS (2001)
Keyphrases
- neural network
- reinforcement learning
- navigation tasks
- collision free
- mobile robot
- motion planning
- obstacle avoidance
- robot navigation
- path planning
- landmark recognition
- visual landmarks
- action selection
- robot control
- service robots
- planning process
- indoor environments
- goal directed
- rough terrain
- unknown environments
- multiple robots
- real robot
- autonomous robots
- autonomous learning
- dynamic environments
- perceptual aliasing
- action selection mechanism
- robot arm
- autonomous navigation
- learning algorithm
- autonomous vehicles
- function approximators
- robot moves
- macro actions
- multiresolution
- human robot interaction
- multi robot
- state space
- visual navigation
- planning problems
- robot behavior
- neural network model
- function approximation
- outdoor environments
- collision avoidance
- image registration
- back propagation
- partially observable
- robotic systems
- robot teams
- world model
- complex domains
- initial position
- deterministic domains
- markov decision processes
- optimal control
- topological map
- blocks world
- coarse to fine