Research On Short-term Forecasting Method Of Traffic Flow Based On Intersection

As the node of urban road network, intersections play a crucial role in the whole urban network system. The intersections on the urban trunk road become the bottleneck of urban traffic in most cases. A little block intersections may trigger the ripple effect on the whole line and even road network. When the intersections of trunk roads are reasonably controlled,congestion and parking delays can be reduced. In this way, traffic efficiency can be improved and environmental pollution can be reduced. And the developed intelligent transportation technology can exactly provide a good optimization scheme for intersection signal control. As an important constituent part of intelligent transportation system, vehicle dynamic induction system has become an effective way for traffic management department to dredge urban road traffic. And real-time and accurate flow prediction is the basis and key to realize the dynamic path induction. Therefore, it is of great significance to study the short-term prediction of urban road traffic flow.Traditional prediction methods just consider spatial relation rather than time factor based on the historical data of single intersections. The paper takes the traffic flow prediction of urban relative intersections as research content. It first analyzes and studies the current situation, development trend and existing problems of the short-term flow prediction of domestic and foreign intersections, simply elaborates data acquisition, traffic flow characteristics and traffic prediction principle included in traffic flow prediction, and then studies the grey prediction model and analyzes and improves its shortcomings.In view of the deficiencies of traditional methods, the paper suggests considering comprehensively time and space and focusing on the research on the ratio of flow paths and travel time of vehicles in the upper and lower intersections. According to the spatial geographical relationship between relative intersections, the paper puts forward the method of the online and rolling prediction of the short-term flow. It is unnecessary to store all the data of traffic volume with the method. In other words, the method ensures prediction continuity, improves prediction accuracy, and reduces error discretization degree. Combined with modern traffic data collection technology, the paper puts forward the conception of predicting and studying the traffic flow of the objectives based on the flow statistics of secondary relative intersections, and verifies the accuracy and feasibility of the model based on the cases of the intersection of Chengdu Fucheng Road-Yizhou Road and the intersection of Fucheng Road-South Chenghan Road.