Research On Short-term Traffic Flow Prediction And Control Based On Big Data Environment

With the rise of artificial intelligence,aiming at the current situation of traffic congestion and traffic pressure surge,researchers put forward intelligent transportation system in the field of transportation.It is necessary for the intelligent transportation system to have a traffic flow prediction system with high accuracy and fast timeliness.Through the accurate and real-time traffic flow prediction system,the traffic flow of the target road in the next 5 minutes to half an hour is predicted.According to the predicted value,the traffic control management system is used in advance to regulate the road traffic conditions,so as to alleviate congestion and improve the traffic efficiency.Therefore,in this thesis,the short-term traffic prediction is more accurate and timeeffective,and the optimization of the performance parameters to reduce the regional traffic delay time is studied.(1)Aiming at the problems of explosive growth,low prediction accuracy and slow prediction speed of current traffic flow data,a combined prediction model integrating multiple factors is proposed.Firstly,K-means + + algorithm is used to cluster Time characteristic factor data to reduce the correlation between different classifiers,and then a prediction model is constructed by using stochastic forest algorithm with strong generalization performance.The experimental results show that compared with the traditional intelligent prediction model,the prediction accuracy of this model is significantly improved.In addition,in order to solve the problem of too long prediction time,this chapter chooses spark technology with distributed parallel processing ability to train the random forest algorithm in parallel.On the premise of ensuring the accuracy of the experiment,the prediction time is reduced to a certain extent.Through the experimental analysis,the overall requirements for accurate prediction and realtime traffic flow are satisfied.(2)In this thesis,the traffic coordination control model with the shortest delay time is established,and the improved delay model is proposed for the Webster model in the regional traffic coordination control without considering the saturation state of road flow and the HCM2000 model without considering the phase difference factor between internal roads.The traffic flow in the area is divided into internal road and external access road.The coordination mechanism of vehicle delay and phase difference is analyzed.All intersections in the road network area are associated to establish the regional traffic coordination control model.(3)In addition,based on the vehicle delay model,the improved genetic algorithm is used to optimize the period,phase and phase difference of each intersection in the region.In the algorithm,the probability of crossover and mutation is associated with the average fitness and individual fitness of the population,which improves the speed of the solution.In the process of regional optimization,some optimal individuals of each intersection are combined into the initial population of regional calculation,so as to further improve the convergence speed of the algorithm.(4)Finally,a four intersection area is modeled and simulated by using the VISSIM simulation software.The field sampling data,the results of traditional coordination delay model and the results of optimization model proposed in this thesis are input for simulation operation,and the delay time of each intersection is obtained.The experiment shows that the scheme in this thesis can effectively reduce the delay time of vehicles in the area.