The Partition Of Urban Traffic Network And Classification Of Traffic Status Based On Clustering

In recent years, with the improvement of urbanization level and the increase of car ownership, a series of problems including traffic congestion, travel safety, and environmental pollution gradually emerge. The way of simply increasing the urban infrastructure not only costs manpower and material resources, but also has limited effect. Therefore,how to manage and control traffic effectively is an important research content of Intelligent Transportation System (ITS).Urban traffic network is a nonlinear time-varying system which is complex and large in scale. It is difficult to conduct a unified way of management and control. After the urban traffic network partition, it is possible to carry out targeted control program for each sub-region, so that the road network system becomes flexible and reliable. In addition, the amount of data that needs to be processed in real time is significantly reduced, and the real-time requirements of the road network system can be met.In this paper, Wangjing region is selected as the research site, and the VISSIM simulation platform is set up. Based on Beijing floating car data, the relevant parameters of the simulation platform is set, and the traffic environment of the study region is reproduced. The clustering algorithm is used to classify the Wangjing region, and the partitioning results are evaluated from the qualitative and quantitative views by using the Macroscopic Fundamental Diagrams (MFD).The main contents of this paper are as follows:Firstly, considering the velocity and density information only, the K-means clustering algorithm is used to partition the road network. And then the spatial location information of the link is added to the algorithm and partition the network again. The two methods are compared. The clustering results show that the proposed method is more effective. And the partition effect of K-means algorithm is evaluated by MFD of each sub-region. And through the fitting function of MFD, a standard for measuring partition effect is proposed.Secondly, the improved fuzzy C-means (FCM) clustering algorithm is used to determine the clustering class number and the initial clustering center, then Wangjing area network is divided. The partition effect of improved FCM algorithm is evaluated by MFD of each sub-region. The results of the evaluation show that the partitioned results of the improved FCM algorithm are more ideal than the partition results of the K-means algorithm.Finally, the fuzzy comprehensive evaluation method is used to determine the traffic state of the links in sub-region. The congestion degree of sub-region is changed with the time, and the evaluation criterion of sub-regional congestion is proposed by using the nature of MFD. The dynamic partition of Wangjing area network is analyzed,and the change of traffic congestion with time is analyzed, and the key road sections are determined.