Metric State Space Reinforcement Learning for a Vision-Capable Mobile Robot
Viktor ZhumatiyFaustino J. GomezMarcus HutterJürgen SchmidhuberPublished in: CoRR (2006)
Keyphrases
- state space
- reinforcement learning
- mobile robot
- reinforcement learning algorithms
- markov decision processes
- optimal policy
- path planning
- topological map
- action space
- computer vision
- markov chain
- heuristic search
- obstacle avoidance
- vision system
- robot control
- dynamical systems
- state abstraction
- dynamic programming
- state variables
- partially observable
- learning algorithm
- metric space
- markov decision process
- particle filter
- autonomous navigation
- control problems
- mobile robotics
- motion planning
- indoor environments
- real time
- function approximation
- dynamic environments
- search space
- model free
- metric learning
- real robot
- multi agent
- continuous state spaces
- mobile robot navigation
- autonomous robots
- unknown environments
- state and action spaces
- action selection
- macro actions
- reinforcement learning methods
- motion control
- search algorithm
- learning process
- distance metric
- partially observable markov decision processes
- planning problems
- robotic systems
- belief state