Research On Oversaturated Queuing Overflow Based On Electronic Police Data

As the contradiction between vehicle growth and road capacity is becoming more and more intensified,traffic congestion is often caused.During the morning and evening rush hours,the phenomenon of road congestion overflow occurs when it is severe.When this phenomenon occurs,the traffic police department is often required to launch police force for signal timing or manual command adjustment,which is inefficient and wastes police force.Therefore,in order to solve this problem,intelligent transportation is urgent.In order to improve the traffic capacity of vehicles during peak hours,adjust the overflow caused by road queuing congestion,improve road carrying capacity,and realize refined queuing management,research on queuing overflow at oversaturated intersections,and establish a process based on electronic police data.Queuing overflow model at saturated intersections.In order to distinguish the different traffic flow states that may be presented by the same travel time,the vehicle travel time,delay,departure and arrival time obtained by the electronic police license plate recognition technology are used as the input parameters of the model,based on the delay of the bicycle and the queuing relationship between the vehicles,using The traffic flow theory and the vehicle flow dissipation model,combined with the comparison of the green light cycle emissions at the intersection and the vehicle capacity of the road section,established the overflow interval model of the short-circuit section and the long road section to determine whether the vehicle has the possibility of overflow.On this basis,the research object of the first vehicle in the queue during the overflow period is introduced,and the relationship between the departure time of the first vehicle and the overflow vehicle in the overflow period is used to establish a discriminant model to further determine whether the vehicle overflows on the two road sections.Finally,according to the START urban traffic health diagnosis system,the actual intersections that fit this research are screened out,combined with the microscopic traffic simulation software,and the accuracy of the VISSIM verification model is established.In order to prevent queuing overflow at oversaturated intersections and compensate for the lag of electronic police data.First,based on the classification ideas of decision trees and random forests in machine learning,overflow prediction modeling is carried out,and the data of VISSIM simulation experiments are used for verification.In order to improve the accuracy of the prediction model,after comparing various models of the neural network,the LSTM neural network is selected to construct an overflow prediction model with vehicle travel time and overflow conditions as inputs,so as to achieve the effect of predicting and preventing vehicle overflow in advance.