Research On Key Technologies Of Semantic Service-oriented IoT Middleware

The vision of the Internet of Things is that more and more things are embedded in the capabilities of recognition,perception,driving,processing and communication,which penetrates all aspects of society through wide application and provides users with simple and convenient intelligent application services.Since the development of the Internet of Things,most of the application systems have been developed in a closed and tightly coupled manner,while the access of massive complex and diverse devices makes IoT with scale and heterogeneity characteristics.These lead to the independence of application systems from each other and cannot well integrate and share resources,resulting in low resource utilization in physical and network environments,which hinders the further development of IoT.The current scope of relevant research on IoT is limited to vertical application services such as sensing and identification,but the horizontal direction of cross-domain and cross-system convergence applications can provide more intelligent services and realize the vision of IoT in a real sense.Therefore,it is necessary to build the Internet of Things middleware to change the current vertical fragmentation development mode of the Internet of Things and support IoT infrastructure integration,aiming to combine heterogeneous devices and technologies into application development and provide a unified standard IoT services access interface upwards to achieve data interaction and sharing between different applications by shielding the complex details of the underlying implementation.The ability of IoT to connect to a large number of smart devices gives it certain advantages,but it also brings new challenges.The high degree of differences in devices,computing communication capabilities,protocols and data formats make the main challenge in the development of IoT is the heterogeneity of data and functional operations,which causes problems such as data not being understood by each other and information not interacting properly between system applications.This makes semantic interoperability one of the key issues to be addressed by IoT middleware in providing seamless communication and services.Semantic models describe data in a machineunderstandable processing mode,which can enhance the value of data information and help exchange semantically annotated information between heterogeneous applications.The service-oriented idea can abstract the functions of sensors and other devices in things with a unified and standardized operation model as services,so that heterogeneous sensory semantic data can be accessed through the Internet services to achieve the overall interoperability of diverse applications of the Internet of Things.At the same time,with the extension of the application field,the number of connected device entities scale and diversification of device types,IoT data and service space continues to expand.The IoT application goal of providing relevant data and personalized services according to user needs makes the efficient distribution and delivery of data services also become a key issue to be solved by IoT middleware.In order to solve the above problems caused by the scale and heterogeneity of IoT,the main innovative work of this paper on IoT middleware and its key technology research content are as follows:1.Constructed a semantic service-oriented IoT middleware modelAiming at the problems of non-uniform format and heterogeneous interoperability of IoT devices,based on the analysis of semantic interoperability in IoT environment,this paper uses the idea of semantic and service computing model to build a semantic service-oriented IoT middleware model.The middleware model integrates semantic and service technologies to support an information semantic interoperability environment between the sensing device layer and the application layer.2.Ontology-based semantic modeling of the Internet of ThingsTo address the problems of "understanding gap" arising from data heterogeneity among IoT underlying devices and the inability to interact and share information.By analyzing the mapping relationship between the physical world and the information world,this paper establishes an Ontology-based semantic sharing model of conceptual entities in the heterogeneous IoT environment,which mainly includes the ontology models of device-resource and function-service.The model describes sensory data in a machine-understandable way,adds semantics to IoT data association,and abstracts device functions into services for unified specification description,which lays the foundation for information interaction and sharing among cross-domain and crossplatform system applications.3.On-demand distribution message space based on publish/subscribe modelFor the huge message space formed by the rapid data acquisition in the IoT environment,which causes the problem that users cannot get valuable information quickly and directly.This paper construct a message space for on-demand data distribution in IoT middleware based on the publish/subscribe interaction model.In order to focus on solving the problem of reduced performance of the event-subscription matching algorithm caused by large-scale user subscriptions,a novel subscription index structure based on frequent attribute sets is designed using the relevance of user interests and the two-level subscription filtering strategy reduces the size of the subscription set for subsequent matching.A corresponding event matching algorithm based on interest priority is proposed,which uses atomic predicate constraint and subscription predicate constraint range as the basic operations to match eventsatisfiable subscription sets.The design of this algorithm filters almost all the subscriptions that are not related to events,removes a large number of duplicate or similar predicates from the subscription set,and reduces the number of redundant comparisons in the matching process.The experimental results show that it can significantly improve the performance of the event-subscription matching algorithm.4.Similarity-based IoT service discovery methodIn response to the large number of devices connected to the Internet of Things(IoT)and the wide range of application domains involved make the number of services provided grow continuously,leading to difficulties for users to obtain services that meet their needs,this paper constructs a unified service provisioning management space by integrating IoT services ins IoT middleware.In order to improve the performance of service discovery provision in the context of IoT,a service probabilistic discovery method based on semantic similarity is proposed.The method firstly divides the service discovery space using probabilistic clustering of service topics oriented to domain ontology,and reduces the number of service functional logic similarity calculations by filtering the similarity between the topics of requested services and clustered topic clusters.Then the semantic similarity of service functional logic is calculated using the information content of ontology concepts to match and filter services that functionally satisfy user needs.Finally,non-functional services satisfying users are discovered based on quality of service Qo S and user preference weights.The method reduces the spatial quantity scale of service discovery by probabilistic topic similarity,and discovers IoT services that fully satisfy users' needs using functional logic similarity and quality of service Scores.The performance is relatively better than some other service discovery methods through the analysis of the experimental results of accuracy,F1-measure,average recall and precision.