Ontology-based Service Retrieval And Invocation Mechanism For Services In Smart Home

The smart home service refers to the service that can be implemented on thehousehold appliances. By integrating the services provided by different appliances, itcould assist advancing the intelligence of smart home, and could facilitate appliancesmanaging remotely by employing controller or other devices which could accessInternet. However the standards adopted by the appliance manufacturers are suchdifferent in technical layer. As lacking a uniform standard, the services which aredeveloped with different standards could hardly communicate with each othereffectively.SOA is a well adopted architecture for Web Service organization. It provides aseries of standards for service description and communication. All the SOA standardsare platform-independent and Extensible Markup Language (XML)-basedspecifications. With the assistant of SOA, software agents could know how to accessthe services which are described by such standards, and could invoke and composethe services with instruction of such standards as well. Therefore labeling the homeservices with the standards provided by SOA could uniform the access interface of theservices, and as a result the services developed with different protocol couldcommunicate with each other conveniently. However the standards of SOA areexpressed by XML which does not provide the concepts like class and property,therefore the standards lack enough well-understood formal semantics for agentsinterpretation. So, actually, the service discovery, selection and invocation are basedon keywords matching. Similar to other web sources retrieving, the keyword-basedservice searching approach has several drawbacks, including problems with synonyms(semantically similar but syntactically different services) and homonyms(syntactically equivalent but semantically different services).OWL-S is an OWL-based ontology that provides a semantic markup languagefor specifying Web Services in an unambiguous and computer-interpretable form. In conjunction with domain ontologies for smart home field, it could provide standardmeans of specifying declaratively descriptions for home services that enable selectionand invocation automatically. However, the domain ontologies proposed in literatureare not sufficient. There still three problems at least exist in the ontologies proposed inliterature. First of all the classification of services lacks a fixed viewpoint, i.e. theyfail to form an upper ontology to describe the relations among concepts formally.Secondly the classification of semantic of services has some deficiencies as well.Thirdly the content which is consumed or produced by service and device which hostsservice play important roles in smart home, unfortunately, these concepts are absent inthese ontologies.In order to solve the above problems, this paper focuses on the following fivetopics for home services.(1) This paper presents domain upper ontology named home service ontology forsmart home. The ontology explicates our viewpoint and formally definesservice-related concepts from the unified viewpoint. With such viewpoint theservices in our paper would have fixed meanings which are independent withcontexts. Then, by refining the concepts in home service ontology, an ontologicalmodel for home service is given to annotate the semantics from different facets.(2) By analyzing the relations among the concepts of home service ontology, thispaper defines the function as the semantic of service, and constructs functionconcept ontology for service classification. In the function concept ontology,firstly the concepts of function are divided into five categories by analyzing thesatisfaction process of need. Then the concept of function in specific category isclassified into sub-concepts based on the realization ways.(3) This paper proposes content ontology to annotate the semantic of content that theservice operates. The concepts in the content ontology are extracted by analyzingthe effect of content in information expression. With the content ontology, wecould filter the service semantically and also could differentiate the services withsame functions to optimize service searching(4) This paper presents device ontology to provide the semantic information for device. Since the implementation of service still relies on special device,specifying the device information could assist service retrieval within a givenhome environment. On the other hand in the home environment, a contextreference list is employed to record the device information provided by theenvironment. By comparing the information recorded on the context reference listwith the instance of device ontology, appropriate services could be downloadedfor personalized home environment.(5) This paper proposes function decomposition tree to assist service developersanalyzing the functions of composite service. Then a temporal functiondecomposition tree, which is an expanded function decomposition tree, is given topresent the construction and sequence within the functions of composite service.Finally we translate the temporal function decomposition tree into OWL-S Modelin order to facilitate agents parsing. The atomic processes in the model areannoated with domain ontologies from ontological model in order to composeservices semantically.In this paper, a scenario of audible alarm of gas detective is given to present thearchitecture for service registry, retrieval, invocation and composition based on theontological model. Finally a prototype system is developed to illuminate theeffectiveness of our approach.