A proposal for self-recognition in robot programming

In this paper we propose a way to perform robotic self-recognition in static and quasi-static environments, self-recognition is a process during which the robot discovers the effects of its own actions on the environment and itself. For example, how much would the robot move when its wheels turn once? Such information can be hard-coded into the robot\´s code, but it can also be discovered automatically, self-recognition can be used when the robot or the environment are unknown a priori or change. Examples are when new wheels or a completely new robot is used, or when obstacles are added to the environment. Instead of re-programming, the user can let the robot "play" and discover how to change its situation. In this way only the self-recognition phase should be repeated, while the high level planning strategy can remain the same