Abstract

We propose that the development of causality can be seen as a primitive for understanding and constructing complex systems either biological or artificial. Furthermore, we put forward a view of development in terms of the control of complexity. Although some of these elements are at the moment speculative or barely outlined, the theoretical test and verification are part of the ongoing research. On the artificial side, we will show how developmental principles are used within the architecture of a humanoid robot. The reference problem is the ontogenesis of sensori-motor coordination. Visual, acoustic and inertial cues constitute the sensory repertoire of the robot; computation, in the form of mappings, represents its brain activity. The continuous and meaningful adaptation during the natural interaction of the robot with the environment is one of the key aspects of the implementation.

Citations