Study On Beijing Expressway Microscopic Traffic Simulation Model Based On HD Video Processing Technology

Intelligent Transportation System (ITS) integrates effectively the advanced information, data communication transmission, electronic sensor, electronic control and computer processing technologies. And a complete integrated transportation and management system are constituted of ITS. This system plays a role in a wide and full range and operates in real-time, accurately, and efficiently, in order to ease the traffic congestion problem. The prerequisite or basis of ITS to play a role is to obtain the basic traffic parameters, such as road traffic information (traffic volume, vehicle type, speed, vehicle trajectory).With the economic and social progress of our country, the transportation industry is also developing rapidly. Analysis and research for road traffic flow in line with China’s actual situation, has become an important research direction in the transportation field of China. Simulation means of microscopic traffic flow is an important tool for study in this field.In this paper, vehicle video detection technology is used to continuously extract the information of the position coordinates and speed of vehicles, and synchronize the collected information real-time to the information database. Standards-based video is obtained by a series of pretreatment such as video editing. On the basis of the standard video, the extraction of vehicle trajectory data is completed by using the Vehicle Tracker. In this paper, oversaturated traffic flow model for express roads is designed, and the parameters are estimated by combining measured data such as vehicle trajectory data. Finally, based on particle swarm optimization (POS) method, the method of parameters calibration for microscopic traffic simulation model is discussed, and examples demonstrate the effect of the method.The first chapter briefly introduces the research background, purpose and significance of research and technical routes.The related researches at home and abroad in the second chapter are combed, including the current research of traffic video detection technology, the current research of the car-following and lane change model. The third chapter describes the video pre-processing process. In order to extract the static and dynamic information of vehicle from videos, collected from high-performance cameras, a series of complicated pretreatments, including video editing, video image stabilization processing and the video correction, are needed for videos. A variety of software such as VirtualDub, Steady Hand is used in the whole process. After a series of videos processing, the original videos have been become standards-based videos continuous-time and continuous-space on the observed sections.The fourth chapter describes the Vehicle Tracker software. It obtains the position of the vehicle position from the videos shot by the camera, and convents them to the vehicle tracking data. This chapter first introduces the main technical points of the Vehicle Tracker; and then the four windows’(the main control panel window, view window, warning window and warned the editor window) functions are described in detail. It also describes the database used to store Vehicle Tracker output data, and provides a brief description on how to manipulate the database. Finally, examples illustrate the applications of the Vehicle Tracker.The fifth chapter introduces activity algorithms of microscopic traffic simulation model for oversaturated traffic flow about express roads. The method is a comprehensive car-following model and lane changing model framework, and in accordance with the kinematic-wave theory. This algorithm explicates mandatory and selective lane changing model, including cooperative lane changing. As the extension of the method, a new on-ramp junction model is put into it. The algorithm contains a little number of parameters, and it is easier to achieve on-site measurement.Chapter six describes the estimations of parameters including car-following and lane changing models in oversaturated traffic flow model about express roads. The model parameters are extracted from the Vehicle Tracker data, which are obtained from the loop test points in Deshengmen bridge. This chapter describes the estimation methods, and results of car-following parameters and lane changing parameters are given, and it also provides a comparison of model estimation results according to the test point.In chapter seven, the method of traffic simulation model parameter calibration, which is based on particle swarm optimization method, is designed. First, the calibration process of the simulation parameters is developed, and then the microscopic traffic simulation software VISSM is selected as a platform to establish a simulation model parameter calibration algorithm based on particle swarm optimization (PSO) algorithm; the main parameters effecting the simulation results are analyzed and calibrated, and automated calibration of VISSM simulation parameters is realized. The method is applied to the driver behavior parameter calibration of Beijing highway simulation model.Finally, the main conclusions of this study have been sorted out in this paper, and the future research directions are prospected.