Research And Implementation Urban Area Traffic Prediction And Control

With the the new urban construction in my country has made outstanding achievements,and people’s living standards have been continuously improved.However,due to historical reasons,the infrastructure in some areas of the city was not able to meet the growing travel needs of urban residents.A number of problems such as traffic congestion,frequent traffic accidents and car pollution have become chronic diseases in cities,and traffic controversies are becoming more visible.The resolution of the above contradictions,based on the perspective of managing urban traffic lights,is one of the effective methods generally recognized by domestic and foreign scientists.Under such circumstances,this thesis optimizes the coordinated management of regional traffic flow through signal management,conducts an in-depth study of a short-term urban traffic flow prediction method,proposes a traffic separation method,and implements hierarchical regional traffic.signal management on this basis.Based on the existing traffic simulation engine,a set of experimental traffic simulation systems have been developed and implemented to support hierarchical regional traffic signal management,which has a profound effect on reducing a number of congestion problems caused by adverse traffic.control.The main content of the research and innovations of this thesis are as follows:（1）This thesis proposes a CNN-Bi LSTM short-term traffic forecasting model.For the LSTM-based short-term traffic prediction model,there is a problem that the characteristics of the road network structure cannot be fully generalized.This model uses the idea of a convolutional neural network to connect with it to capture the spatial eigenvalues.traffic data and then pass the Bi LSTM model.Spatial traffic data is extracted using time series functions to complete short-term traffic flow prediction and compared with traditional CNN neural network,Bi LSTM and classical prediction algorithm.ARIMA,it shows the accuracy and reliability of the model in this thesis.（2）This thesis proposes a traffic sub-district division scheme.The separation scheme is based on the CNN-Bi LSTM short-term traffic flow prediction prediction results,calculates the similarity of each intersection flow and signal cycle similarity according to the predicted traffic flow,and separates each traffic sub-region in detail,and sets the subregions in combination with the traffic flow prediction time interval.Dynamically adjust time interval,and then implement dynamic adjustment of the traffic sub-area according to the accumulated "refresh index".A multi-level control scheme for the regional coordination of traffic lights based on the division of transport sub-areas is proposed.The scheme is based on division into subdistricts and separates the regional coordinated traffic management to simplify complex management tasks.（3）This thesis designs and implements a virtual simulation experiment system that supports the coordinated control of hierarchical regional traffic signals.The regional coordinated control scheme is assembled in the existing simulation system,and the overall architecture and main functional modules of the traffic simulation experimental system supporting hierarchical regional traffic signal coordination control are designed and implemented.