Urban Traffic Signal Control Research And System Design Based On Traffic Flow Prediction

In recent years,with the number of vehicles in urban has been increasing,which has brought about the rapid increasing in road traffic pressure.The traffic congestion problem has aroused great concern.On the one hand,road intersections,as the basic unit of urban traffic signal control,are the key nodes for alleviating traffic jams.On the other hand,traffic signal control is the most effective way to manage traffic at intersections.Therefore,according to the analysis of domestic and foreign research review,this paper focuses on the research of urban traffic signal control based on traffic flow prediction.The main research contents are as follows:(1)The traffic flow prediction model based on deep learning is studied.First of all,the spatial-temporal correlation and the characteristics of rainfall impact of traffic flow data are analyzed.According to the influence of rainfall,the improved K-means algorithm is used to cluster the historical traffic flow data.Then this paper proposed a novel forcast model based on deep long short-term memory(LSTM)neural network combining the spatial-temporal characteristics of traffic flow data(ST-LSTM).Besides,the historical traffic data clustered by K-means algorithm is used to train the prediction model in the Tensorflow framework,the future traffic volume is predicted by the trained model.By comparing with the ARIMA,BP and SVM prediction models,the experimental results show that the ST-LSTM prediction model has higher prediction accuracy.(2)An intersection adaptive signal control model based on multi-objective optimization model is proposed.A multi-objective evolutionary algorithm is used to solve the optimal parameters of the multi-objective optimization model chosen CO2 emissions,vehicle cycle delays,traffic capacity and cycle stops as optimization objectives.According to the short-term traffic flow prediction data and traffic conditions detected at intersections in real time,this paper designs an adaptive signal control algorithm for intersections.Finally,by simulating in VISSIM4.3 traffic simulation software for one practical intersection,the simulation result shows that the adaptive control method can effectively reduce CO2 emissions,vehicle cycle delays,cycle stops and increase the traffic capacity of vehicles at the intersection compared with the timing control and induction control methods.(3)A signal coordination control method based on the algebraic algorithm for bottleneck intersection in the arterial road is proposed.This paper employs the algebraic algorithm to analyze the arterial road signal coordination examples and put forward the problem that the green wave bandwidth of arterial road signal coordination is seriously restricted by the bottleneck intersection.For this kind of arterial road with bottleneck intersections,the bottleneck intersection is equivalent to straight road when the distance between the bottleneck intersection and the adjacent intersections satisfies some certain conditions,then the algebraic algorithm is employed to solve coordinated signal control parameters for the equivalent arterial road.Furthermore,in order to respond to the coordination control signal from adjacent intersections,the bottleneck intersection dynamic control model is established with maximizing the traffic capacity of bottleneck intersections.Through the simulation in VISSIM software,it's effective for the bottleneck intersection signal coordination control method proposed in this paper.(4)The signal coordination control of multi-intersections based on multi-agent coordination is studied.Each intersection is regarded as an agent,information can be exchanged between adjacent agents.Based on the agent's ability of storage,memory,expert learning,reasoning,and forecasting,a multi-intersection coordination signal dynamic programming model is proposed.Finally,Through VISSIM-COM programming simulation,compared with the MAXBAND model,the results show that the proposed dynamic signal coordination control approch can greatly improve the coordination control effect.(5)Under the framework of the traffic cloud platform,based on the Hadoop cloud computing platform and the Nansha District intelligent traffic control platform in Guangzhou,the subsystems of traffic flow forecasting system,isolated intersection signal adaptive control system and multi-intersection signal coordination control system are designed,which is implemented on Nansha District intelligent traffic control platform.