Unguided Robot Navigation Using Continuous Action Space.
Sirinart TangruamsubManabu TsuboyamaAram KawewongOsamu HasegawaPublished in: ICANN (3) (2010)
Keyphrases
- robot navigation
- action space
- continuous state
- state space
- continuous state spaces
- markov decision processes
- real valued
- continuous action
- reinforcement learning
- autonomous robots
- stochastic processes
- control policies
- scene understanding
- action selection
- landmark recognition
- markov decision process
- vision system
- dynamic programming