Navigation Method Selector for an Autonomous Explorer Rover with a Markov Decision Process

Simon Le Gloannec (1) , Abdel-Illah Mouaddib (1)

1 Equipe MAD - Laboratoire GREYC - UMR6072

Many high level navigation methods exists for autonomous rovers. Usually, fixed methods is choosen for the rover to follow its trajectory. In some cases, a rover can have several embedded navigation methods, but there is currently no automatic approach to select the best method. In this paper, we propose an automatic navigation method selector using Markov Decision Process (MDP) framework. Navigation method are modelized in the MDP transitions functions. Results we achieved were promissing.

Domaines: Intelligence artificielle [cs.AI]; Modélisation et simulation; Système multi-agents [cs.MA]

Simon Le Gloannec, Abdel-Illah Mouaddib. Navigation Method Selector for an Autonomous Explorer Rover with a Markov Decision Process. Proc. 2nd International Conference on Intelligent Robotics and Applications (ICRA~2009), 2009, Kobe, Japan. ⟨hal-00965859⟩