Autonomous Decision And Planning Of Robot Service Based On Ontology In Intelligent Space

Service robots are gradually accessing to our daily life,playing an increasingly important role in family,medical,office and other fields.People put a higher demand on the quality of robot service,they hope the robot can provide intelligent and personalized services autonomously.For service robot,how to acquire full knowledge and understand the needs of users based on the knowledge are keys to improve the service capabilities.Intelligent space technology can assist the robot to acquire space knowledge and has made some achievements in the practical application.But the diversity,heterogeneity and dynamism of resources in space make it difficult for the robot to realize the integration and sharing of various knowledge.Therefore,how to organize the knowledge effectively,improve the expressive ability of intelligent space knowledge,and provide support for the robot service recognition and execution becomes an important research content.To solve the above problems,the paper designs an ontology-based cognition and planning method to help the robot to provide services automatically.The main contents are as follows.Firstly,according to the intelligent space resources,we establish an ontology model based on ontology technology which can eliminate the limitations of intelligent space.The ontology domain includes three parts:the environment domain,the user domain and the service domain.Objects in space can be matched in specific field.The intelligent space ontology library will be formed by storing the ontology model in relational database,and it can realize the update of knowledge.Secondly,Jena is used to extend the knowledge in the ontology library.The custom rules are established to mining space hidden logic knowledge,user-centered adaptive data-concept conversion mechanism is established to map data knowledge to concepts which can represent preferences of users or qualities of services.Based these,the intelligent space ontology library can drive more knowledge based on the original information,provide more complete information for robot service inference.Thirdly,based on the extended ontology library of intelligent space,service inference rules in SWRL for robot can be constructed.JESS inference can match the ontology library with rule base to generate the service sequence which the robot needs.Knowledge related to inference results can be extracted to generate the problem description of JSHOP2 planner.JSHOP2 planner can complete the planning of service tasks by using the hierarchical task network planning algorithm.Finally,the method proposed in this paper is applied to robot which developed by Service Robot Laboratory in Shandong University.The experimental results verify the effectiveness and real-time of the proposed method.