Urban Road Network Traffic Volume Prediction Based on Road Section Speed

摘要

With the development of urban construction and the growth of vehicle ownership and traffic demand, the status of urban transportation networks has changed a lot. In order to re-examine the characteristics of urban traffic flow, this paper learns from the Greenshields model, and uses the observation data of intersection and the road section speed estimated by Baidu API to establish the speed-density, density-flow relationship model of each road segment. Then, considering the traffic flow characteristics of the road network, the Transcad OD Matrix estimation is used to reverse the data, and establish the corresponding road section speed-flow model. By comparing the differences of regression coefficient, the road section formula with no significant difference is merged. Finally, the heteroscedasticity of the model is tested, and it is considered that the elimination of heteroscedasticity can optimize the model.