Design And Implementation Of Highway Traffic Flow Prediction System Based On Cloud Platform

With the expanision of the highway network and the expansion of people’s demand for cross-border travel,the problem of expressway traffic congestion is becoming more and more serious.Based on the traffic history data,traffic flow forecast contributes to realizing traffic guidance in advance,which is a hot research topic in the field of intelligent transportation.Traffic state is difficult to be predicted.The traffic flow prediction algorithm in the traditional serial computing mode can not work efficiently.It is difficult to achieve the prediction of traffic flow between prediction accuracy and computational efficiency,facing massive traffic data.The development of cloud computing is helpful for processing massive data.This paper,based on cloud platform,realizes a distributed and parallel freeway traffic forecasting algorithm on the premise of efficient storage management of massively charged data,and completes the prototype system of highway traffic forecasting.The main work of this paper is as follows:Firstly,according to the characteristics of highway toll data,proposing a hierarchical data warehouse model which helps to store and manage massive traffic data.Secondly,establishing a highway traffic forecast model based on BP neural network.And improving the BP algorithm by cuckoo algorithm in terms of training.In order to overcome the shortcomings of the traditional single-node computing model,improving the computational efficiency of the algorithm with the help of Spark,when dealing with massive data.Finally,completing a highway forecasting system which contributes to highway traffic flow forecast and multi-dimensional analysis.It is proved that this system is valuable and practical.Experimental results show that Spark parallel prediction algorithm performs well.