Identifying The Mode Of Bus And Car Using GPS Data

In recent years,with the acceleration of urbanization,the number of motor vehicles has increased rapidly,causing serious traffic problems,which hinders the healthy development of urban construction.To make reasonable traffic planning and management policy on the basis of resident travel survey data and analysis results are vital measures to solve the problems,so to obtain reasonable travel survey data becomes important.The traditional resident travel survey depends on artificial investigation,which is greatly influenced by the subjective cognition of respondents.In the investigation,some trips are often forgotten or slipped,which reduces the accuracy of the data,and we could not obtain the specific travel information,such as travel path and location.Obviously,it could not meet the needs of current transportation research and management both in accuracy and quantity.GPS devices are widely used in resident trip survey due to the advantages of convenient portability,accurately recording travel time and space location.Although the application of GPS devices avoids some abuses in the traditional travel survey,it proposes great challenges to process GPS data and fetch feature information from the data.Identifying travel mode is an important issue in the research of travel behavior analysis,and it is difficult to identify bus and car owing to their similar operating characteristics.To solve the problems provides data for the analysis of trip characteristics.Therefore it is of great significance for us.This paper presents the techniques of detecting bus and car based on GPS survey data,according to the rules of classification and machine learning algorithm,considering the GIS information as well as the impact of traffic conditions.Firstly,design an algorithm to extract feature parameters of the travel trajectory and velocity,combined with the GIS information,such as bus lines,bus stations and road network.Then set the rules to identify bus and car,simultaneously build a support vector classification mode to detect them.Besides,it considers the influence of traffic state to the results.comparing with the actual travel mode,the results indicate that the algorithm based on the support vector machine has a better accuracy,and the recognition accuracy of non-congested state is higher than that of congestion state.Finally,test the transferability of the two methods,and the results show that the second method is more applicable to other cities.The methods to identify bus and car in this paper is feasible and effective,and it is a supplement to the research of travel mode detection.The results can be used to the extract travel information and analyse behavior based on GPS survey data,which will significantly improve the efficiency and quality of survey data.In addition,it is of great significance to promote the application of GPS technology in resident travel field,reduce energy consumption,promote the healthy and sustainable development of urban traffic system.