Research On Threshold Control Of Urban Road Network Based On Short-Term Traffic Flow Forecasting

Rapid economic and social development has led to a significant improvement in people’s living standards and an increase in the number of cars in cities.Urban road infrastructure cannot meet the increasing traffic demand,and the imbalance between supply and demand makes the urban congestion problem increasingly serious.Therefore,how to effectively alleviate urban congestion has become an important topic of modern intelligent transportation research.In this thesis,the research results at home and abroad are analyzed in depth,and the regional threshold control strategy of urban road network based on short-term traffic volume prediction is studied comprehensively.Firstly,based on the microwave coil detection data of urban regional road network provided by Open ITS open platform,the advantages,disadvantages and applicability of the current traffic information collection methods are counted and discussed respectively;and the characteristics and spatio-temporal correlations of the traffic data after pre-processing are analyzed,as well as the specific reasons for this phenomenon;meanwhile,the close connection between traffic parameters and traffic status is clarified.Second,a combined prediction model based on 1DCNN+LSTM+Attention is constructed to fully exploit the temporal well and spatial characteristics of urban traffic flow data.The spatial characteristics of traffic flow are first captured using 1DCNN,and then the temporal characteristics of traffic flow are analyzed using LSTM to obtain as many patterns of historical traffic flow as possible,so that the characterization parameters of traffic data can be accurately predicted.In order to verify the feasibility of the proposed model,the sample data are first input into the model to obtain the predicted values of traffic,and it is found that the predicted values match well with the observed values,which indicates that the prediction accuracy of this research model is high and the usability of the model is confirmed,and it also shows more intuitively the process of traffic state change in the future time of urban road network,which helps to release traffic information in advance.Finally,in terms of threshold control,a threshold control strategy based on macroscopic fundamental graph theory is proposed.The short-time traffic flow prediction data are input to the threshold control model,combined with genetic algorithm to optimize the data of road network input traffic,road network output traffic and road network output traffic to construct a threshold value based on the maximum cumulative number of vehicles that the road network can accommodate,and study the maximum cumulative number of vehicles that the road network can accommodate according to the macroscopic road network map;meanwhile,a Vissim-based simulation model is constructed with the existing road network as an example,and the traffic demand planning to test the effectiveness of regulation.The simulation results show that the implementation of the threshold control strategy significantly improves the incoming and outgoing traffic flow of the road network,which enables the road network to maintain a high capacity;at the same time,the travel time of the road network vehicles is reasonably controlled,and the average speed and average delay are also improved to some extent.The above research results show that,for traffic congestion prediction and congestion control problems,the combined model of short-time traffic flow prediction and threshold control strategy proposed in this thesis can solve the adverse effects caused by traffic congestion to a certain extent,which has important theoretical and practical values.