Traffic Flow Analysis Using Real-Time Traffic Data

With the continuous development of urbanization, traffic problems have become increasingly serious. Using Intelligent Transportation Systems (ITS) to service urban traffic is an effective way to solve the problems. Traffic State Monitoring System (TSMS) is one of the most important subsystems of Intelligent Transportation System (ITS). Real-time information obtained by TSMS is the basis of congestion management, traffic control, route guidance and so on.Global Positioning System (GPS) has been widely used in taxis and other vehicles. So that it is possible to collect real-time traffic information using GPS-equipped probe vehicles. The information of position, instantaneous velocity and so on could be received from them. Over the past few years, many researchers have proposed several methods to estimate traffic conditions by integrating GPS and Geographic Information Systems (GIS). Two of them are the Curve-Fitting Estimation Method and the Vehicle-Tracking Estimation Method.Curve-Fitting Estimation Method set up a traffic velocity model using the velocity distribution regularity of traffic flow on one certain road link. And the least square method is used to solve the parameters of the model. Using the fitted model, we formulate an improved average velocity of traffic flow which is based on both time-varying and space-varying. Furthermore, the required minimum number of probe vehicles that are necessary to value real-time traffic flow state for a desired statistical accuracy is investigated. Consequently, a sample size algorithm is developed to calculate the minimum sample sizes of this method. Meantime, road style, road segment length and sample frequency are taken into account in the algorithm.The Vehicle-Tracking Method collects traffic data with GPS-equipped probe vehicles and the A* heuristic search algorithm is employed to judge the optimal vehicle tracking path. In addition, the average velocities of road links are calculated by considering the velocities of vehicle tracks as well as their corresponding credibility factors.In order to guide the application of the two methods, we carried out some experiments with massive real traffic data of Shanghai to compare the performance of them. In the experiment, data from thousands of GPS-equipped taxies were taken as the probe vehicles, and the estimate accuracy and operation speed of the two different methods are systematically measured. For the accuracy experiment, the ground truth was obtained by repeating the videos which were shot on 24 links in Shanghai downtown. The experiment results illustrate that the Vehicle-Tracking method usually bears higher estimate accuracy and slower operation speed, compared to the Curve-Fitting method.