A traffic-network-model-based algorithm for short-term prediction of urban traffic flow

In the research field of Intelligent Transportation Systems (ITS), traffic flow prediction is a key technology for traffic guidance and advanced control strategy. Accuracy and immediacy are the main requirements for prediction methods. This paper presents a short-term prediction algorithm of traffic flow rate based on the macroscopic urban road network model. Classified into different typical elements, a traffic road network can be expressed as a matrix. Taking crosses and their links as basic research objects, the proposed prediction method can only use a few real traffic parameters obtained from loop detectors to realize accurate short-term prediction of traffic flow rate. This method can also be adaptable to different kinds of road network. In case study, the real traffic system is simulated with the microscopic traffic simulation platform (CORSIM). In the given simulation environment of road network, the experiment results illustrate that the proposed prediction algorithm can accurately predict flow rate in short term.