Study On Analysis Of Urban Road Network Traffic With Applications

Traffic flow analysis,urban road traffic state comprehensive evaluation and forecasting have become the basic and critical work in the development of intelligent transportation system.It can provides great help not only for traffic management department in grasping the overall operation of road network and analyzing the causes of traffic congestion,but also for travelers in advoiding traffic jams.The research has very important theoretical significance and application value to improve the quality of traffic management and service.Oriented to the actual demands of traffic management and public travel,the theory and application of critical section identification,urban road traffic state evaluation and long-term forecasting have studied deeply in this paper based on the traffic data.Then,a relatively complete road traffic state analysis system is formed.At last,this paper verifies the feasibility and practicability of the findings through application of experiments.The main innovative work of this paper is embodied as follows:(1)Based on the spatial and temporal characteristics of traffic flow,the space-time correlation analysis of traffic flow is proposed,and it is also applied to identifying the critical sections in the road network.Using space-time correlation function as the index,the influence of traffic state between different sections is analysed from the two aspects of space and time.In the framework,spatial weight matric is introduced to explain the adjacency relation of traffic flow and the structure of road networks.Taking the road networks as whole,the ST-ACF measures the inuence of traffic flow on network performance with some time and space lag,while CCF measures the strength and direction of inuence between a certain section in the road network and its neighbors from a local view.In addition,the application of CCF to deal with the effect of a section on others has also been used to explain the importance of section in road network.Based on technique for order preference by similarity to an ideal solution,the model is constructed to identify critical section.The experiment shows that the proposed model is practical and feasible,and can identify critical section effectively.(2)Based on the traffic flow information obtained from the detectors,this paper establishes a multi-hierarchies evaluation index system of urban road traffic state to meet the needs of different traffic participants.According to the multi-level demands of traffic management,traffic state evaluation and travel information service,the principle and structure of evaluation indexes are determined,and this paper establishes the evaluation index system of urban road traffic state.It describes the change mechanism of traffic state from both time and space in terms of three levels:the microscopic section,the mesoscopic road and the macroscopic road network.At last,according to the structure of road network,the definition and calculation of each evaluation index for section,road and road network are explained respectively.(3)Based on the established evaluation index system of urban road traffic state,a comprehensive evaluation model is proposed using fuzzy entropy-entropy weight mothod.This model can be divided into two parts.Firstly,considering the complexity and fuzziness of traffic flow,the improved FCM clustering analysis is introduced to classify traffic flow into 6 different states:very smooth,smooth,stable,slow,congested and very congested.The aim of FCM method is to determine the fuzzy range of each index in different states.Secondly,the fuzzy range,obtained by the improved FCM method,is transformed into intuitionistic fuzzy numbers using fuzzy entropy theory.As the input of the comprehensive evaluation model,the intuitionistic fuzzy numbers are used to compute the weights of each evaluation index by entropy weight mothod.It is worth noting that importance of section and product of road volume and length are employed as the weights of section and road in the comprehensive evaluation model.In the framework,the evaluation results of traffic state are also computed corresponding to section,road and regional road network.The experiment shows that the findings of traffic state evaluation are in accordance with the actual situation and people's sense of traffic state.It can meet the demands of traffic management and service,and combines organically the qualitative description and quantitative analysis.(4)Based on the seasonal pattern present in space-time correlation of traffic flow,a new long-term traffic state forecasting model is introduced by kernel estimate of nonparametric regression.The method is to estimate the the probability distribution of traffic state in the future through the evaluation results(comprehensive evaluation index),using nonparametric regression function.It consists of three steps:First,according to local space-time correlation(CCF),the influence of its neighbors on forecast target is calculated for selecting the state vector.Second,the functional principal component analysis is used as distance function for computing proximities between traffic flow curves in different periods.Third,the proposed model is applied to Beijing expressway for long-term traffic state forecasting by selecting kernel function and the optimal window width.The experiment shows that the model outperforms other forecast methods in accuracy and effectiveness.Complementary,it is of universality and expansibility,and can be easily applied to other evaluation indexes in different levels whenever they become available.(5)Combining the theoretical research and practical application,the analysis of application case for Beijing road traffic information service is put forward.Based on the actual needs of Beijing traffic management department and travelers,two kinds of traffic information service are proposed corresponding to two different traffic participants.One is traffic information evaluation and statistics service oriented to traffic management,the other is traffic information prediction service oriented to public travel.Meanwhile,this paper devises the corresponding application scheme of traffic information publishing service.Combined with the practical application case in Beijing,the paper introduces the contents of traffic information service report in detail,and verifies our theoretical research achievements are of practicability and effectiveness.