Search

Studies intelligence in the context of system-environment interaction. Discussion on the concept of sensory-motor coordination; Application of the New Artificial Intelligence perspective to the problem of categorization; Replacement of the metaphor of information processing by one of sensory-motor coordination.

Abstract: Many different animal species rely on whiskers for a variety of tasks. Among the most basic are obstacle avoidance and wall-following. We show that fast and reliable evaluation of distances can be greatly improved by an appropriate sensory morphology that matches the physical space of the agent. To investigate morphologies and material properties of whiskers, we conduct experiments both on a real robot and in simulation using artificial evolution. We find that the morphology that is most successful at following a wall strongly resembles the morphologies of natural whiskers. This can be better understood by considering the relation of the agent body to its tactile sensing range, as well as the challenges of various tasks faced by agents endowed with different sensory modalities. [Copyright &y& Elsevier]