Estimation of Urban Traffic State Based Velocity Cell Transmission Model

Abstract

The traditional cell transmission model takes the number of vehicles in the sub-sections (cells) as the state variable. In traffic state estimation, the system output equation needs to be established by combining fundamental diagram, increasing the uncertainty of the system and reducing the estimation accuracy. At present, vehicle trajectory data has been widely used in traffic state analysis and optimization, and velocity data is easy to obtain. Therefore, under the current background, directly use mobile device data to build traffic flow model has become an important research problem in the traffic field. Firstly, this paper studies the cell transmission model with velocity as the state variable, which is called the Cell Transmission Model for velocity (for short CTM-v). Further, with network control data and traffic flow parameters as inputs, the CTM-v model is programed by using MATLAB, and then the state evolution of road network is analyzed. Finally, taking the topological network in Changping District of Beijing as an example, the simulation environment is established by using VISSIM simulation software. The results show that the simulation errors compared with CTM-v model evolution are below 5 km/h, except for the large error in the transition of congested and non-congested state, which validate the accuracy of the proposed CTM-v estimation model in this paper.