Urban Traffic Environment Partition Based On Cluster And D-S Theory

The urban traffic system is refers to the city's technical level in each aspect of transportation, the management strategies, the rules and regulations as well as the municipal transportation custom and so on. The traffic pollution question and the traffic congestion question affect people's daily life in varying degrees. The urban traffic system is extremely complex, and a city's transport network is also quite huge, therefore carries on the unification to the entire urban traffic, the intelligent control and the optimization is not realistic. Therefore, the theory of traffic division is proposed. Some scholars divide the entire transportation system into different small regions, and these different small regions have the same attribute, then through carries on the intelligent optimization to the transportation subsegment to achieve to the entire transportation system's government and the optimization.The paper is based on consults the domestic and foreign articles of traffic division, summarize of predecessors'research results and profits from other people experience. Through thorough analysis of cluster method and the municipal transportation system, the cluster method is applied in the traffic division, we attempts to divide the traffic map according to the similarity of evaluating indicator data. The main contents of this paper are as follows:We consider the factors of the different characters which the transportation system status distributed in spatial, sources of data, operability and so on, correspondingly evaluation indexes are selected; First, we transform the whole transport system into many smaller grids; each lattice has n feature indices that represent transport properties, each n-dimensional traffic data could be seen as a pixel in image. Because after the non-division grid's traffic network can regard as an image, so the method of image division can be used in the traffic division. In view of characteristics of urban traffic, method characteristic and the usability of the ISODATA algorithm and grey clustering have analyzed, and improved them, then these two methods used in the traffic division. Finally, D-S evidence is adopted to solve uncertainty information that generated in the process of classification.