Research And Application Of Key Technologies Of Ubiquitous Objects Unified Monitoring And Control

The development of Internet of Things technology has brought us into the era of Io T when “things” and human are interconnected,and has also promoted the upgrade of traditional monitoring systems.However,due to the heterogeneity existed in various aspects of the Internet of Things,the current Io T eco-system is still in a relatively fragmented state,forming a relatively closed system from clients to devices.Many Io T platforms lack a unified semantic description of devices,and the data generated by the devices form information silos,which are difficult to be utilized by other devices,resulting in waste of data value.At the same time,it is difficult to make the devices cooperate together.Developers need a huge amount of work to integrate various devices to build a unified monitoring application.In view of the aforementioned problems,in this thesis,we study key technologies of ubiquitous object unified monitoring,which aims to simplify unified monitoring application development by abstracting objects monitored into "ubiquitous objects",enhancing semantic interoperability with unified description model,and provide related frameworks and tools to simplify the development of unified monitoring applications.The work carried out here is as follows:(1)In view of the lack of semantic interoperability of current Io T platforms,and combined with the requirements of unified monitoring,a ubiquitous object model is proposed,and a unified monitoring framework based on the model is designed.The semantic layer in the framework is used to achieve the semantic unification of data.In order to simplify the development,the concept of virtual ubiquitous object is proposed,and an action mapping method of virtual object based on inference utilizing the semantics in ubiquitous object model is designed,which achieves the goal of adaptively selecting appropriate devices to perform the action of a virtual object according to the environment;(2)The key technologies of the data processing layer in the unified monitoring framework are studied.The semantic annotation module is designed and implemented,connecting the data processing layer and the semantic layer,and transforming the data into the representation of knowledge;two typical data processing modules: the event processing module and the anomaly detection module are designed and implemented.Esper event processing engine is used as the core of the event processing module and Event Processing Language(EPL)is used to define events.An improved anomaly detection algorithm based on time series decomposition is used in the anomaly detection module.(3)To reduce the difficulty of developing a monitoring application,we use Vue.js framework to implement a configurable monitor client.Based on the componentization development idea,the data structure of various dashboard components is designed.The dashboard components are encapsulated and provided to developers in form of the “component library”.Developers can design the monitoring interface through visual “drag and drop” operation,and can flexibly define the monitor interface and monitoring logic.Finally,the effectiveness and practicability of the monitoring platform were verified by demonstrating the functions of the unified monitoring platform.