OntPercept: A Perception Ontology for Robotic Systems

摘要

In the last few years, the need to use robots in unstructured environments in direct contact with humans has led to the growth of Human-Robot Interaction (HRI) area, especially social robotics. This article focuses on the approaches used to process the sensory information in a highly complex and dynamic world. Initially, the relevance of defining a formal communication protocol for sensory information is emphasized by the necessity of exchanging information among the modules of a robotic cognitive framework. Next, a cognitive ontology called OntPercept is proposed for modeling the sensory information necessary to perform the activities involved in HRI tasks. Thanks to the ontology model proposed an additional gained, that being the access to the structured information of the Semantic Web technology. This proposal has been validated using a cognitive simulator specifically designed for this purpose. The main contribution of this study is minimizing the development costs through secure and reliable communication among the components of architectures designed for social robotics.