Meta-Context Ontology Based Context Representation And Context Reasoning Research In Semantic Web Of Things

Internet of Things(IoT)makes various IoT-devices be networked.Under IoT situation,people can achieve inter-things connections via various sensors and communication technology,and can get various sensors' context information,so that human perception is extended but massive heterogeneous information is produced.And,things in IoT are evolving from the materialized ones to abstract ones,which expand the connotations in IoT.Information system contains six basic elements including human,abstract things,materialized things,machine,network,time,and their inter-relations.From this perspective,to solve the problems of Structural Holes in social network and human-machine mutual incomprehension,semantic technology and context knowledge about Semantic Web and Pragmatic Web should be integrated into IoT,so that Semantic Web of Things(SWoT)occurs as a brand-new IoT,which can establish the semantic associations among different information islands and fill the gap among information flows in information exchange to achieve semantic interoperability.Based on Semantic Web,IoT and social network,SWoT can build an open world service ecosystem that constructs and integrate physical space,social space and cyberspace,which can fill the gap between them by discovering and using context information to solve the heterogeneity in IoT development and achieve Semantic Collaboration.Hence,to achieve SWoT's goal and interoperability and to.solve inter-information data friction,it is necessary to deal with the heterogeneity,dynamicity and uncertainty of data in SWoT,so a general and extensible meta-context ontology model is proposed in this paper to unify the modeling and representation and to avoid the heterogeneity for all dimensions on information space.To balance dynamics and uncertainty,a Dynamic Bayesian Network based dynamic uncertainty reasoning algorithm is proposed to achieve the uncertainty reasoning.By inferring the high-level contextual state on its subject from probabilistic and temporal dependences of context information,then its adaptive contextual services can be obtained.Finally,related experiments have been done to verify and evaluate the feasibility and effectiveness of the architecture and the prototype system proposed in this paper.