【作者】 王清波；

【导师】 钟珞；

【作者基本信息】 武汉理工大学 ， 计算机应用技术， 2013， 博士

【摘要】 随着城市建设和发展需要,隧道作为缓解拥塞的交通要道,数量逐年增加,其隧道运营安全日益引起人们的重视。本论文将以城市公路隧道监控为对象,研究通用的监控系统架构、异构多数据源的集成、联动控制方案以及交通数据智能分析方法等关键技术,从而建立一个集隧道运营监测、交通控制、灾害预警报警等功能于一体的较为通用的城市公路隧道智能监控系统,其主要工作包括如下几方面：首先,根据武汉市多条隧道监控系统的设计和实施经验,总结和分析了城市公路隧道监控系统的通用架构、子系统划分及基本功能、实现方案及技术手段。提出了一种多个子系统分散控制和集中管理的体系架构。然后,研究并建立了一种城市公路隧道的交通态势本体模型,采用语义本体技术建立城市公路隧道交通态势本体模型和知识化表示。进一步提出了一种本体集成的通用框架,它在语义网中用全局本体来提供一个统一的本地本体视图查询,该框架描述了一种设计空间,可用来解决语义网应用程序中本体集成的问题。接着,分析和整合各个监控子系统的数据和功能,提出了基于XML的各类数据的统一表示和存储方法,探究了基于数据挖掘的多源数据融合的联动控制机制,并建立了隧道预案管理平台。更主要的是,研究了隧道交通数据的特征,以最小二乘法的数据压缩方法为基础,提出了一种针对隧道交通数据的核主成分分析(KPCA)两步降维算法,以约简数据；研究了隧道交通数据流的分类和预测方法,以时间段划分训练样本,提出了一种基于RBF核函数的支持向量回归(KSVMR)堵车情况预测模型。根据武汉市多条城市公路隧道监控项目的实际需要和测试结果,本文提出的通用城市公路隧道监控架构及关键技术,能较好地解决实际问题,保证了隧道的安全环境,舒适行车和有效管理。更多还原

【Abstract】 With the development and construction of the modern city, the number and quality of the urban highway tunnels which have become the traffic arteries easing congestion are improved. It is necessary to pay much attention to these tunnels’ monitoring system and operational safety. This thesis proposes a common tunnel monitoring architecture, integration of heterogeneous multisource, the linkage traffic control scheme and traffic data intelligent analysis methods in order to build an appropriate monitoring system integrating running monitoring, traffic control, disaster warning and maintenance and management and so on. The main works of this thesis are as follows.Firstly, according to our design and implementation experience of multiple tunnel monitoring systems in Wuhan, urban highway tunnel monitoring system’s general architecture, subsystems division and basic functions, implementation and technical methodologies are summarized and analyzed. A new mechanism of multiple subsystems decentralized control and centralized management is proposed.Secondly, the tunnel traffic trend ontology model and knowledgeable representation reflecting basic concepts and semantic model of integrated transport network trend are researched and the established. A general framework for ontology integration is further proposed which used in the global semantic web ontology to provide a unified view of the local check. The framework describing a design space can be used to solve the semantic web ontology application integration issues in order to support integrated transport network management and improve the whole efficiency.Thirdly, various monitoring subsystem data and functionality are analyzed and integrated. XML-based unified data representation and storage methods for all kinds of data are proposed. The multi-source data fusion and data mining linkage control mechanism is explored. At last, the tunnel planning management platform is established.Last but not least, the characteristics of the tunnel traffic data are studied. On one hand, a new Kernel Principle Content Analysis (KPCA) for tunnel for traffic data using two-dimensionality reduction algorithm on the basis of the least square method is proposed and tested to reduce data. On the other hand, the tunnel traffic data stream classification and prediction methods for time-division training samples based on Kernel Support Vector Machine Regression (KSVMR) model for the traffic jam situation are proposed and verified.On the basis of practical requirements and testing outcomes of Wuhan urban highway tunnel projects, the framework and techniques proposed in this thesis are able to solve the practical problems efficiently. The aim of this thesis is to build a common intelligent urban highway tunnel monitoring system in order to ensure the safe environment, comfort driving and efficient management.更多还原