Prediction And Applied Research Of Short-term Traffic Flow Based On Elman Neural Network

In recent years,with the rapid development of economy and continuous prosperity of cities in our country,the per capita car ownership is growing day by day,so the large,medium,or small-sized cities are confronted with various degrees of problems such as traffic congestion,traffic safety,energy consumption and environmental pollution,which exert high pressures on the quality of urban life and sustainable development of the urban economy.Accelerating the construction of urban transportation system and highly developing the intelligent transportation have become a common voice of those cities in the new era.As the basic and crucial link to realize the intelligent transportation control and guidance,the real-time and accurate short-term traffic flow prediction and its application research have an important theoretical value and guiding significance to improve the current flowing efficiency of transportation network and to solve the present urban transportation problems.The paper takes the short-term traffic flow prediction and the city intersection traffic signals' optical control as objects and puts forward the short-term traffic prediction methods and the traffic lights polling control model on the basis of IPSP-E1 man neural network and threshold service strategy respectively.Finally,the simulation result shows that the short-term traffic flow prediction accuracy and the traffic flow capacity of urban intersection can be improved effectively.The paper's main points include:In accordance with the highly nonlinearity,randomness,correlation and other features of the short-term traffic flow,after comparing and analyzing the advantages and disadvantages of the traditional and new intelligent predicating methods,the paper chooses the E1 man neural network with highly nonlinear mapping ability,good learning ability,and dynamic change ability as prediction method,and combines the improved particle swarm optimization algorithm with the E1 man neural network to complete each other so as to better speed up the traditional neural network convergence and avoid falling into the local minimum.In the paper,the short-term traffic flow prediction model based on the IPSO-E1 man neural network and adopt IPSO algorithm to optimize E1 man neural network's connection weights and threshold is built,through the analysis of periodic and change law for the collected data of traffic flow in time and space,the specific input of the prediction model is determined.The simulation experiment shows the prediction precision of the short-term traffic law has been improved.Based on the prediction results' research and application of the short-term traffic flow and the vehicle volumes' correlation analysis of the urban traffic intersection,in the paper,the optical control method rested on the urban intersection traffic signal of the vehicle volumes is put forward,traffic lights polling control model is build,the traffic light timing project and the average queue length of vehicle crossing the intersection are figured out.Finally,the simulation experimental result shows that the intersection flowing efficiency can be improved effectively and the traffic congestion can also be alleviated very well.