TY - CHAP

T1 - Travel-mode classification for optimizing vehicular travel route planning

AU - Zhang, Lijuan

AU - Dalyot, Sagi

AU - Sester, Monika

PY - 2013/1/3

Y1 - 2013/1/3

N2 - Navigating and travelling between destinations with the help of Geographic Information Systems route planning is a very common task carried out by millions of commuters daily. The route is mostly based on geocoding of the addresses given by the traveller based on static road network into digital-map positions, and thus the creation of path and directions needed to be taken. Today’s navigation data sets rarely contain information about parking lots, related to building entrances, and walking paths. This is especially relevant for large building complexes (hospitals, industrial buildings, city halls, universities). A fine-tuned route tailored for the driver requirement, e.g., park the car close-by to destination, is required in such cases to save time and frustration. The idea of this chapter is to extract this information from the navigational behaviour of users, which is accessible via an analysis of GPS traces; analysis of car commuters in relation to their point of departure and destination by analysing the walking path they took from—and to—their parked car in relation to a specific address. A classification scheme of GPS-traces is suggested, which enables to classify robustly different travel modes that compose a single GPS trace. By ascribing the classified vehicular car trace, which is accompanied by a walking path to/from the car, to a specific address, it is made feasible to extract the required ascribed data: parking places corresponding to that address. This additional data can later be added to the road network navigation maps used by the route planning scheme to enable the construction of a more fine-tuned optimal and reliable route that will prevent subsequent detours.

AB - Navigating and travelling between destinations with the help of Geographic Information Systems route planning is a very common task carried out by millions of commuters daily. The route is mostly based on geocoding of the addresses given by the traveller based on static road network into digital-map positions, and thus the creation of path and directions needed to be taken. Today’s navigation data sets rarely contain information about parking lots, related to building entrances, and walking paths. This is especially relevant for large building complexes (hospitals, industrial buildings, city halls, universities). A fine-tuned route tailored for the driver requirement, e.g., park the car close-by to destination, is required in such cases to save time and frustration. The idea of this chapter is to extract this information from the navigational behaviour of users, which is accessible via an analysis of GPS traces; analysis of car commuters in relation to their point of departure and destination by analysing the walking path they took from—and to—their parked car in relation to a specific address. A classification scheme of GPS-traces is suggested, which enables to classify robustly different travel modes that compose a single GPS trace. By ascribing the classified vehicular car trace, which is accompanied by a walking path to/from the car, to a specific address, it is made feasible to extract the required ascribed data: parking places corresponding to that address. This additional data can later be added to the road network navigation maps used by the route planning scheme to enable the construction of a more fine-tuned optimal and reliable route that will prevent subsequent detours.

KW - Classification

KW - Data mining

KW - GPS

KW - Optimization

KW - Route planner

UR - http://www.scopus.com/inward/record.url?scp=84988500382&partnerID=8YFLogxK

U2 - 10.1007/978-3-642-34203-5_16

DO - 10.1007/978-3-642-34203-5_16

M3 - Contribution to book/anthology

AN - SCOPUS:84988500382

T3 - Lecture Notes in Geoinformation and Cartography

SP - 277

EP - 295

BT - Lecture Notes in Geoinformation and Cartography

ER -