【作者】 马骏；

【导师】 朱跃龙； 冯钧；

【作者基本信息】 河海大学 ， 计算机应用技术， 2006， 硕士

【摘要】 随着社会的发展,交通问题日益突出,人们对智能交通系统投入了越来越多的关注,其中道路网模型作为智能交通系统的基础成为研究热点。目前,在大多数的研究中,道路网作为空间网络被抽象为图模型。考虑到道路网的丰富特征、属性和复杂性(如道路的转弯限制等交通约束),常规图模型已不能满足要求。因此,出现了结点连接模型和伪二重图模型。这些模型以增加图中结点和弧数量为代价表示部分交通约束(如转弯限制),使得基于图算法的空间查询效率降低。 针对上述问题,一种超点模型提出将交通约束集中在各结点上,并以约束矩阵形式表示。本论文在完善超点模型基础上,进行了以下几方面的实现:1、使用扩展的UML工具对道路网进行建模,利用图形来描述真实世界各个对象的符号表示。2、给出了与概念模型匹配的数据模型,通过融合代价弧与约束矩阵的标识符减少了数据冗余。3、对超点模型支持基于道路网的查询算法结合空间数据索引进行了研究。 论文最后对超点模型和结点连接模型进行了存储空间、查询支持等方面的模拟实验与比较。结果显示,对同一道路网的表示,超点模型的存储开销小于后者;对连续最近邻查询和区域查询,在超点模型上的实验效率均高于后者。表明超点模型不仅对道路网和交通信息有着良好的表示,还能提高基于道路网的查询效率。更多还原

【Abstract】 As the problem of transportation is increasing today, the intelligent transportation system is arousing more and more attention. Being the basic of the intelligent transportation system, the road network model becomes the focus of the research. At present, road network as a spatial network is abstracted to be a graph in most of the researches. Considering the usual graph cannot meet the demand of the ample features attributes and complexity of the road network, there appear node-link model and pseudo-dual graph model. These models can describe the part transport restriction, but the increment of the node’s and the arc’s quantity makes the cost of the graph algorithm increase greatly.To solve the problem above, a super-node model is presented. This method concentrates the turn restriction on each node and uses restriction matrix representation. Based on perfecting the super-node model, this thesis made some implement as follows: 1. The extent UML instrument is used to model the network, the symbol representations of the real world objects are described with graphic. 2. A data model suited to the conceptual model is proposed. Data redundancy is reduced by merge cost-arcs and the node identifications of constraint-matrix. 3. Some of the query algorithm based on the road network is mended with integrating the research of spatial data index.Simulation experiments on storage space and query support are processed on super-node model and node-link model in the last of this thesis. The results show that the storage cost of the super-node representation method is less than that of the latter, and the Continuous nearest neighbor search and range query on the super-node model is more efficient than the latter. The experiments indicate that the super-node model can not only represent the road network and transportation information well, but also lead to an efficient query process.更多还原