GEOMETRIC AND SYMBOLIC REASONING FOR MOBILE ROBOTICS

Abstract

Autonomous mobile robots such as planetary rover need task and path planning abilities in order to fulfill their assigned missions. Task planning is caracterized by a symbolic reasoning and aims at defining a sequence of actions which will be executed to achieve the goals of the mission. Path planning allows to find some ways in the environment to reach these goals and corresponds to a geometric reasoning. Coupling these two kinds of reasoning presents open issues such as the description of the environment and the consideration of geometrical constraints that must be verified in order to act and move. This paper adresses these issues by studying different ways to couple a symbolic and a geometric reasoning and by proposing a hybrid planning architecture allowing to solve problems in which a mobile robot has to act and move in the environment.