Title :
Classifying trip characteristics for describing routine and non-routine trip patterns
Author :
Millonig, Alexandra ; Maierbrugger, Gudrun ; Favry, Eva
Author_Institution :
Mobility Dept., Austrian Inst. of Technol., Vienna, Austria
Abstract :
Public transport companies need to exhaust the utilisation capacity of their services in the most efficient way. Therefore, especially during off-peak hours, it is necessary to attract more passengers by offering specific, customised services. As part of the scientific project NRT (“Non-Routine Trips”), we examine habits, preferences and interest profiles of lifestylebased user groups in order to identify key requirements for enhancing public transport services and motivating passengers to also use the services for non-regular journeys. During the multidisciplinary approach, we used a combination of different complementary methods in order to gain comprehensive knowledge about the yet unexplored concept of non-routine trip behaviour. This paper focuses on the classification of trip characteristics based on single or multiple trip purposes. We collected 530 trip datasets from self-administered trip diaries completed by 23 volunteers during one week of data collection. For clustering, we used a variant of the family of spectral clustering in order to build clusters of combinations of trip purposes. The clusters have been subsequently analysed according to the volunteers´ specifications concerning further trip characteristic such as regularity or transport modes used for each trip. The results describe patterns of trip purposes and related characteristics reported by the participants, and give first insight into interrelations between trip purposes and regularity of trips. Although the given sample size only allows limited interpretation, additional analyses of personal attributes of participants related to the different clusters obtain promising results concerning the socio-demographic characteristics of user groups showing specific patterns of trip behaviour.
Keywords :
behavioural sciences computing; pattern clustering; traffic engineering computing; nonroutine trip patterns; public transport companies; routine patterns; spectral clustering; trip behaviour; utilisation capacity; Cities and towns; Clustering algorithms; Companies; Educational institutions; Interviews; Medical services;
Conference_Titel :
Intelligent Transportation Systems (ITSC), 2010 13th International IEEE Conference on
Conference_Location :
Funchal
Print_ISBN :
978-1-4244-7657-2
DOI :
10.1109/ITSC.2010.5625222