Are shaping techniques the correct answer the control of visually guided autonomous robots?

Abstract

Motor Category first learning phase corridor following reflex Figure 8: A two levels perception-action association architecture, RMO1 is used by RMO2 as if it is a sensory output. Each level corresponds to the basic perception-action association architecture. 4 Navigation in an open environment The navigation in an open environment may a priori seem also more complex. Indeed, if we keep the approach developed in section 2 and 3, the problem is not here to find a single object, but a set of local views relevant for place recognition. This entails that the robot should "understand" the objects in its environment (for instance it must know if objects can move). Yet, an animal introduced in a room can learn how to reach a goal easily even without a complete understanding of all the objects. The PerAc architecture can be used to learn the association between a perceived place (represented by a set of local views and their angles in an absolute or a relative referential) and the direct...