The Analysis of Characterization of Urban Traffic Congestion Based on Mixed Speed Distribution of Taxi GPS Data
The taxi with GPS is an efficient measure for detecting traffic condition. It is often called as floating car or moving detector. The aim of the paper is to estimate the characterization of urban traffic congestion based on taxi GPS data. Owing to the various factors including signal control, heterogeneous driver behavior, various vehicle performance, speed distribution of urban traffic is the typical mixed distribution. Based on this understanding, the paper firstly used kernel density estimation technique to estimate the probability density of mixed speed distribution. This method was a non-parametric probability density estimation method. Under the precondition that Gaussian kernel obtained the good fit quality, the paper used mixed Gaussian model to analyze the characterization of the congestion. By mixed Gaussian model, the paper obtained the numerical index including the mean, variance, weight. The example shows that we can estimate the characterization of urban traffic congestion using the paper's method. The results are important for designing traffic management plan for different scenarios and evaluating the performance of traffic management facilities.