Urban Road Traffic Condition Prediction Algorithm Based On Bus Big Data And Its Application

With the rapid development of the social economy,the problem of urban road congestion has become increasingly prominent,which has caused great trouble to the daily life of urban residents.As an important part of the intelligent transportation system,the traffic condition prediction technology can accurately identify the traffic state of the road section and provide beneficial guidance information for the traffic participants,and effectively alleviate the traffic congestion problem.At present,various sensor technologies are gradually maturing,and a large amount of traffic data is accumulated.Mining useful information from massive traffic data has become a serious challenge for building intelligent transportation.Based on the analysis of the accuracy and practicability of the existing traffic condition prediction algorithm,this paper proposes an urban road traffic condition prediction algorithm based on public transit big data and its application.The main research includes the following aspects:(1)In view of the fact that the existing methods for calculating the average speed of sections do not take into account the influence of the unstable running state of buses in and out of the stations on the speed calculation,a new method for calculating the average speed of sections based on dynamic weights is redesigned.At the same time,the parallel design for calculating the average speed of sections is completed on Hadoop platform.Finally,the improved method and the traditional speed integration are implemented from ARE,MRE,MARE and MAX(ARE).Comparing the algorithms,the experiment shows that the result of the improved calculation method is the closest to the real value of the road speed.(2)Aiming at the problem of using unique data set and single model in existing traffic condition prediction algorithm models,a prediction algorithm model combining time series and artificial neural network is proposed.The algorithm models real-time data and historical data respectively by time series,and uses artificial neural network to adjust the prediction value of real-time data and historical data.The experimental results show that the model can control the prediction error rate within 12.2%,and can effectively predict traffic conditions under different input parameters.Aiming at the problem of poor parallel computing ability of existing traffic condition prediction algorithms under massive traffic data,Hadoop technology is combined with the traffic condition prediction algorithm proposed in this paper to quickly mine the real traffic condition rules from the massive traffic data and improve the operation efficiency of the prediction algorithm.(3)Develop the prototype of urban road traffic condition prediction system and design the flow of the system.Its main functions include real-time road condition display,future road condition prediction,historical road condition query and road condition data analysis.