An Unsupervised Approach for Knowledge Construction Applied to Personal Robots

Abstract

The employment of personal robots or service robots has aroused much interest in recent years with an amazing growth of robotics in different domains. Although sophisticated humanoid robots have been developed, much more effort is needed for improving their cognitive capabilities. Interactions with humans and/or with other agents are still limited and not considered satisfactory. So, the way we store and represent knowledge in a cognitive architecture is fundamental in order to overcome these limitations and improve the human-machine and machine-machine interactions. In this article, we propose an unsupervised approach for knowledge construction based on the robot's perception. Our approach makes use of Kohonen maps as an unsupervised machine learning technique and allows the definition of semantic clusters from visual features perceived by the robot. Besides, a multimedia graph knowledge base using a pure formalism is presented, which can be actively used by personal robots in their classic activities, such as environment exploration or information gathering, to represent and share the acquired knowledge, linking it to abstract concepts gifted with semantic relations.