Semantic Robot Memory Store using 5W1H for Service Tasks

摘要

Nowadays, many researches use the ontology concept to build the knowledge system supporting an autonomous symbiotic human-robot. These researches focus on the design of a database schema of storing ontologies, developing a reasoning system based on the rule. The most important goal of using the ontology concept is known to be the context inferencing. In this regard, we focus on designing the robot memory store, called Robot Brain Store with 5W1H, to support the context inferencing for service tasks. The lack of previous researches is as followings: i) not consider the methodical and semantic 5W1H, ii) poor connections between 5W1H and other concepts. To solve these problems, we designed the robot memory model that is the conceptual model for storing the ontology knowledge such as context, object, space, feature, human and 5W1H ontologies. Also, we designed EventsEpisodic RBS which supports the context inferencing to efficiently store and query 5W1H.