The Research And Implementation Of A Scalable Microscope Traffic Simulation Platform

With the development of social economy, the contradiction between transportation infrastructure supply and demand becomes more obvious which has severely restricted the development of the city. In addition, the traffic system is a complex and huge system, which means it is difficult to resolve traffic problems only by theoretical research or experimental methods. In this case, the microscope traffic simulation approach is suitable:it studies on individual vehicles and can simulate the interaction between vehicles and between vehicle and environment, reproduce the true state of traffic flow and show the implementation process of various traffic control strategies. So it becomes an effective experimental tool and explore ways to address these practical problems.In this paper, a complete microscope traffic simulation platform for freeway and urban road transportation network with a high degree of scalability is implemented based on the studying of the domestic and international research on microscope traffic simulation The main research work is summarized as follows:Firstly, the article defines the memory model of road transportation network include Lane/Edge/Road three-tier structure to describe the complex network structure in reality and the connectivity state under traffic rules.Secondly, the paper implements two models used to simulate the behavior of the driver:The car following model (CFM) controls the behavior of longitudinal movements and the lane change model (LCM) controls the behavior of lateral movements. In other words, the CFM calculates the optimal speed when influenced by other vehicles and environment and the LCM solves the lane selecting problem under influence.Thirdly, the paper implements the core component:a microscopic traffic simulator, which regards the state of entire transportation system at a giving time as a frame. The simulator combines the results of CFM and LCM and iteratively calculates each frame in parallel, so as to continuously update the state of whole system.Then, the paper achieves a full life cycle control of vehicle with a total of 3 parts and 7 stages. Besides the driving simulation part, the vehicle initialization part creates vehicle object and initialize properties and a lowest cost path algorithm under the constraint of traffic rules is proposed. The insertion and recover part realizes advanced vehicle control strategy for entering or leaving the network.Finally, as a platform, the paper provides a variety of interfaces and functionality extension mechanisms. And the paper also implements some special containers and a highly efficient event trigger mechanism in order to provide scalability while ensuring the performance.