Ontology based framework of modeling and reasoning for robot context knowledge

Abstract

This paper introduces Multi-layered Robot Context Ontology Framework (MRCOF) for comprehensive, integrated robot context modeling and reasoning for Robot context knowledge. MRCOF consists of five knowledge levels including rules such as a perception level, a geometry level, an object level, a space level and a situation level. Each knowledge level has meta-knowledge, domain-knowledge and knowledge-instance layer. For each knowledge layer, we use a 6-tuple ontology structure including concepts, relations, relational functions, concept hierarchies, relation hierarchies and axioms. The axioms specify the semantics of concepts and relational constraints between ontological elements at each knowledge layer. The rules are used to infer concepts or relations. MRCOF enables to model integrated robot context information from a low level sensor data to high level object, space and situation semantics. With the integrated context knowledge, a robot can understand objects and related context not only through unidirectional reasoning between adjacent knowledge levels but also through bidirectional reasoning among several levels even with partial information. This proposed framework is represented with first-order logic to maintain an integrated uniform representation.