Research On Residents' Trip Hot Routes And Attractive Areas Based On Taxi Trajectory Data

Residents' trip behavior analysis provides an important reference for urban transportation planning, construction, management and policy development. Current research on residents' trip, mainly based on residents' trip survey. Nowadays, residents' trip characteristics occur significant changes. Road traffic situation presents new features. Traditional residents' trip survey become more and more difficult due to the increasingly high cost of obtaining data and traditional survey cannot effectively reflect residents trip behavior in the change of space, distance, and way.In recent years, with the rapid development of GPS and other location equipment and technology, and the related computing environment, it makes it more convenient to obtain the data of the urban residents' trip activities. Taxi is one of the important traffic tool for residents, trip destination is completely determined by the passengers which can well reflect the residents' trip characteristics. Besides, taxi GPS trajectory data is widely distributed with large quantities and easy to collect, so taxi GPS trajectory data is a good source of data for the analysis of residents' trip behavior. Using taxi GPS trace trajectory to identify residents' trip hot routes and attractive areas provides a reference to scientific development of urban transport planning and traffic improvement.In consideration of this fact, this paper took 10287 taxis with the GPS terminal in Chongqing city as the research objects. Through several data processing steps of a large amount of raw GPS data, the effective trajectories which can be used to cluster and Taxi boarding and alighting points were obtained. Then according to the general law of urban residents' trip, this paper divided full days to different periods that are the morning rush hour(7:00- 9:00), noon(11:00- 14:00) and evening peak(17:00- 20:00). Different clustering methods were proposed and applied to find residents' trip hot routes and attractive areas in different time periods, discovering the implicit information of trip behavior, exploring the new way to solve the traffic planning problems such as road congestion, and providing a scientific reference for location-based services.The main contents of this paper include the following items:(1)This paper introduced the method Douglas-Peucher algorithm into trajectory compression and improved the algorithm by considering the vehicle speed named Vehicle Speed Considered(VSC) Douglas-Peucker. Applied it to the taxi trajectory data, the experiments showed that the method effectively reduced storage space and computational overhead, and the valid trajectory still well described the characteristics of the original trajectory.(2)This paper presented a trajectory similarity measurement based on Longest Common Subsequence(LCS), combined DBSCAN clustering method proposed LCS-BASED DBSCAN. Taking into account the special geography of Chongqing city, using the altitude of trajectory to simplify trajectory similarity calculating. By clusering analysis, valid trajectories were obtained. Then LCS-BASED Hot Routes Extraction(HRE) algorithm was proposed to discover residents' trip hot routes in different time periods.(3)In this paper, DBSCAN clustering algorithm for Euclidean distance was used to taxi boarding and alighting points data, discovering residents' trip attractive areas in different time periods.(4)In this paper, we studied a variety of visualization techniques to mining residents' trip behavior. Eventually those hot routes and attractive areas were well displayed.