Ontology-based humanoid cognition and reasoning of object attributes

Abstract

A method to construct a semantic knowledge base based on object attributes was proposed in order to help robots have humanoid cognition and reasoning about object knowledge to assist them to operate objects. Visual attributes, category attributes, physical attributes, affordance, operation attributes were used to describe object knowledge in a systematic and orderly way, and web ontology language was used to model the object knowledge. The rules between attributes were made by semantic web rule language to fulfil the cognition and reasoning about object attributes. A simulation platform was used to verify that the knowledge base can guide the operation of objects. The query results to the knowledge base show that robots can obtain different kinds of semantic information in a unified form.