Architecture And Intelligent Monitor Data Analysis Research On Tunnels Internet (TI)

With the development of information technology in society and the promotion of urbanization in China, people are brought into the multi-sensory, objects-connected digital ocean. We are in the networked world, including the virtual internet and the Internet of Things (IoT). In the complex and wide network, urban transport is an important foundation to support the economic development of China, and its construction and development speed becomes faster and faster. City Tunnels are key and special traffic-aware points of the "Perception of China", however, because of the late development of urban tunnel monitoring of China, so far non-unified norms and standards of urban tunnel monitoring and non-unified decision-support platform of urban tunnel monitoring, the multiple tunnel "Perceptions" can not be achieved. Although the urban tunnel monitoring systems have been established in major cities, but there are no network urban tunnel monitoring. By studying on exploitation of multiple urban tunnel monitoring systems, the relevant characteristics of urban tunnel were summarized and generalized, the idea of Tunnels Internet (TI), as a special case of Internet of Things (IoT) applied in the field of urban transport and intelligent transportation, was proposed in this paper.For the data isolation, non-standardization, data magnanimity, information heterogeneous complexity, tunnel construction and management separation etc. in multiple ubran tunnel monitoring systems. The research and analysis were carried on the architecture, functional hierarchy, data characteristics, data integration and data mining of TI to solve those existing problems. TI will achieve deep-level data sharing and collaborative applications of ubran tunnel monitoring systems, provide services for collaborative emergency management, urban planning and construction, and long-period monitoring, implement co-scheduling and control linkage of different departments, expand applications of IoT technology, and improve the ubran tunnel monitoring. The main contents and innovations of thesis are generalize into three parts:(1) According to the actual needs and related technology development of urban tunnel monitoring system, the idea of Tunnels Internet (TT) was proposed at first, then Tunnels Internet Architecture (TIA) was put forward, and analysis of five-layer model of TIA were carried out; and then details of five functional layers of TIA were analyzed. According to the analysis of data characteristics of TI, classification and heterogeneity analysis of TI data, the corresponding data integration methods were put forward.(2) Intelligent methods and theory of data mining as the main research means, intelligent data analysis of TI were carried out from the status rules and data prediction. On the status rules mining of TI, at first classification model of traffic congestion based on TI was proposed, and then exploitation model of Tunnels Internet congestion association rule (TI-CAR) was established. Related data at the tunnel traffic congestion time as main objects of study, the potential correlations of every tunnels in the TI were explored, and reliable theoretical models and experimental analysis and comparison were carried out to supply basis for the intelligent management of TI.(3) Traffic data stream predictions of TI were studied. Data streams of TI were analyzed, and prediction analysis methods of TI were disscussed. And then two prediction algorithms (TI-GMDH and TI-LS-SVM) on traffic data stream analysis were proposed and discussed. In the end, by experimental analysis and comparison on real data of TI, conclusion was achieved that, compared with SVM prediction, prediction algorithms of TI-GMDH and TI-LS-SVM proposed in the thesis were more suitable for the prediction of TI.