Identifying Trip Purpose Using GPS Data

With the development of social economy and the growing population and motor vehicle ownership,traffic congestion and the resulting environmental pollution and energy consumption have become increasingly serious.It is particularly important to rationalize traffic system planning and management policy development.It is well known that residents' travel surveys are the data base for traffic planning and management.However,the traditional household-resident travel survey has the problems of heavy investigation tasks,low survey efficiency,low survey accuracy,and the inability to carry out multi-day continuous travel surveys.It has been unable to meet the needs of travel data acquisition and analysis at this stage.Due to its simple operation,portability and accurate recording of the traveler's instantaneous speed,departure time,and latitude and longitude information,the GPS device has gradually been widely used in travel surveys.Compared with the traditional household-based residents' travel survey,the GPS-based residents' travel survey has a higher survey accuracy,and can reduce the burden on the respondents.At the same time,it can carry out a multi-day continuous travel survey,so it has a better application prospect.In this context,the research on the extraction of travel information based on GPS data has become one of the research hotspots in transportation field.Because there is no personal and family attribute information,the purpose of travel identification becomes a difficult point in the extraction of various travel information based on GPS data.At the same time,the purpose of travel is the basis for the extraction of other travel information,and it is also one of the important links in the analysis and prediction of travel behavior.Therefore,it is of great significance to study the method of destination identification based on GPS data to promote the overall identification of travel information,promote the widespread application of GPS-based residents travel survey and promote the effective solution of traffic problems.This paper will research on multi-day travel modeling and identification based on GPS data.First of all,it determine the origin-destination points according to the latitude and longitude information of the trajectory points displayed in the GPS data,and infer the trip destination.Then,according to the daily activity chain theory,divide the daily activity pattern of the traveler based on the length and occurrence order of the activity time.Establish the Logit model to classify the daily travel modes as the main trip,early trip,late trip,stop of leaving and stop of returning,and separate the workdays and non-workdays.Take the travel time,land use and other information into account to establish another Logit model to identify the travel purpose of each travel mode on a single day.Finally,establish a habitual fitting model for multi-day travel destination selection to improve the recognition accuracy of travel purposes.The model can be used for multi-day travel destination identification.Comparing the recognition results with the actual travel data,it shows that as considered the inter-relationships between various trip destinations in a day,the method of recognizing the purpose of each trip based on the division of the daily travel mode improves the recognition accuracy of the trip purpose.Furthermore,it establish the habitual fitting model for multi-day and multi-week travel destination selection,improve the recognition accuracy of travel purposes and allow the model available for multi-day travel destination identification.The method of identifying travel destinations based on GPS data will be discussed in this paper is based on GPS data on multi-day trip destinations and daily travel mode information.The research results can be used for residents' travel surveys and information extraction based on GPS devices or mobile phone location systems,which is conducive to replacing the traditional resident travel surveys with advanced survey methods in the big data environment,extending data acquisition methods,data breadth,and improving data processing accuracy.And the results will provide data foundation for traffic management and planning.