How to extract and to exploit vision data for autonomous mobile robots to operate in known environments

To operate in a known and dynamic environment, a robot needs to extract and exploit the vision data. This requires the introduction of two different kinds of a priori knowledge: visual (to extract) and behavioral (to exploit). In this paper we propose a general methodology to extract and exploit vision data. We illustrate it by an application: the software development for the AIBO robots to play soccer. The difficult task is to extract reliable visual information; this is the reason why a priori knowledge is introduced at the level of vision processing. Without reliable vision information it is impossible for the robot to switch to the right behavior, and consequently to play the game.