Research And Implementation Of Bus Arrival Time Prediction Algorithm

Public transport is an important part of the development of urbanization.The developed transportation system is the indispensable basis for solving a series of problems such as the increase of vehicles brought about by economic development,scientific and technological progress.Public transport system combined with embedded technology,Internet technology and artificial intelligence technology,can significantly improve the quality of public transport services.The bus time prediction service system is an important component of the public transport system.It helps the public to plan travel time and attract more people to choose public transport.Thus alleviating traffic pressure and reducing the energy consumption caused by traffic congestion.Firstly,this paper introduces the research background of the bus arrival time prediction service system,and expounds the research results of bus arrival time prediction algorithm.A brief overview of the pre-project,public safety monitoring system is given.Secondly,this pager introduces the machine learning algorithm and distributed computing technology which used in this research.Taking the driving data of bus 607 in Beijing as an example,this paper design the data preprocessing process in order to reduce the GPS acquisition error According to the statistical results of the data,combined with the special traffic environment in Beijing,six kinds of the main factors influencing the bus driving were extracted.And to address the problem of preprocessing computation of large-scale data set,a distributed system of bus data preprocessing is designed.After analyzing the driving characteristics of the bus and extracting the influencing factors,this paper proposes a bus arrival time prediction algorithm based on Random Forest for the six influencing factors.This paper shows the horizontal comparison of different number of random trees in the model.At the same time the algorithm is compared with the historical average model and the neural network model.The experimental results show that the arrival time prediction algorithm based on Random Forest has a low error.In order to solve the training and prediction problem of large-scale sample set,a distributed training and estimation system of random forest algorithm are designed.Finally,the bus arrival time prediction algorithm model is applied to the actual design of the bus arrival time service system,for the public travel to provide line inquiries,vehicle location query and arrival time prediction query three services.