The Research And Implementation Of Semantic Annotation And Reasoning For Device Metadata Towards Semantic Web Of Things

According to Cisco's predictions,there will be more than 50 billions of devices connected to the Internet by 2020.The devices and produced data are mainly exploited to build domain-specific Internet of Things(IoT)applications.From a data-centric perspective,these applications are not interoperable with each other due to the difference in data format and transport protocols.Cross domain,cross platform and cross terminal IoT service are difficult to integrate and innovate.In recent years,the development of the Internet of things has gradually shifted from the traditional IoT to the emerging Web of Things(WoT).WoT solves the heterogeneous in device communication protocol,unifies equipment resources and ability by exposing them as Web API,facilitates the discovery and interoperability of IoT devices in a cyber-physical system.But more applications involving the combination and decision of CPS systems still need a unified knowledge representation of physical resources to eliminate data heterogeneity.This study focuses on building a semantic annotation framework for device metadata to contruct a unified knowlwdge base(KB).The KB will further improve the interoperability of domain knowledge in WoT.The study also designs and implements a knowledge management and application generation system for semantic Web of Things,which is called SWoT4CPS.The study combines knowledge engineering methodology with machine learning algorithm and proposes a mixed KB construction method.The method mainly proposes a generic domain ontology SWoT-O,which reuses,extracts and combines the common concepts in the existing excellent ontology from the WoT domain.The SWoT-O provides the unified upper semantic template for WoT applications.On the other hand,the study extends the KB by using Entity Linking(EL)algorithm,which links devices' structural metadata to corresponding entity in DBpedia,so that domain knowledge base could get access to the rich common knowledge in DBpedia.Meanwhile,aligning metadata and concept in different applications to the same background KB,enhances interoperability of heterogeneous metadata between different WoT applications.A WoT application development system based on the KB and rule engine is designed and developed.It supports the rapid development of WoT applications.A proof of concept equipment diagnosis and automatic control application in intelligent building scene is also provided.