Analyses on Travel Demand Variation Considering …Lower level of %WPSS means driver makes more...
1
Analyses on Travel Demand Variation Considering User’s Choice Behavior on Urban Expressway Using ETC Data* 東京大学 生産技術研究所 大 大口 口研 研究 究室 室 （交通工学） Tawin Tiratanapakhom What is ETC data? Electronic Toll Collection or ETC Data is obtained when a car passes through the ETC gate at the expressway. By analyzing this data, it is possible to observe and understand the behavior of each driver. This knowledge can be used by the expressway operator in making efficient strategies to manage travel demand. Study Area Analysis of Data from Tokyo Metropolitan Expressway Data collected from August 2010 to March 2011 (8 months) Pooled Data versus Panel Data Two data types from ETC and detector data: Daily variability in trip making behavior Conventional traveler demand analysis methods assume that each driver behaves the same way on weekdays. We try to confirm this by analyzing panel data. Choice of Departure Time In general, drivers are aware of their arrival time at destination because of arrival constraint such as work start time. 【～90% of expressway users use ETC system!!】 Pooled data is obtained by putting together all vehicle data over the observation period. If the same vehicle is observed again it will be treated as a different vehicle. Conventional methods use pooled data and assume that driver behavior is stable over time. Panel data , on the other hand, recognizes the same vehicle on another observation. Thus, we can check if there is variation is driver behavior. There are 5 data points and 5 different vehicles! There is 1 vehicle with 5 data points! Day of week analysis Month of year Results show that number of trips each driver makes is not stable for each day of the week. In general, number of trips made by drivers are more regular on Mondays. Lower level of %WPSS means driver makes more regular number of trips. %WPSS %WPSS In general, drivers tend to reduce their regularity during the end of the year. B Be eh ha av vi io or r i is s n no ot t s st ta ab bl le e o ov ve er r t ti im me e ! ! Departure time Drivers are aware of being late than being early on Mondays and Wednesdays. On Tuesday, Thursday and Friday, late and early penalty are slightly different. D Do o d dr ri iv ve er rs s a al lw wa ay ys s d de ep pa ar rt t e ea ar rl ly y/ /l la at te e? ? We set a Reference Start Time (RST) to evaluate if driver is late or early. Buffer time = 95 th Percentile travel time for a trip – Average travel time (Lomax,et. al. 2003) ETCデータを用いた都市高速の交通需要変動分析 * This research was conducted under the collaborative study with Tokyo Metropolitan Expressway Co. Ltd.
Transcript of Analyses on Travel Demand Variation Considering …Lower level of %WPSS means driver makes more...
Analyses on Travel Demand Variation Considering User’s Choice Behavior on Urban Expressway Using ETC Data*�
東京大学 生産技術研究所 大大口口研研究究室室 （交通工学） Tawin Tiratanapakhom�
What is ETC data?�
Electronic Toll Collection or ETC Data is obtained when a car passes through the ETC gate at the expressway. �By analyzing this data, it is possible to observe and understand the behavior of each driver. This knowledge can be used by the expressway operator in making efficient strategies to manage travel demand.�
Study Area�
Analysis of Data from Tokyo Metropolitan Expressway�Data collected from August 2010 to March 2011 (8
months)�
Pooled Data versus Panel Data�
Two data types from ETC and detector data:�
Daily variability in trip making behavior�
Conventional traveler demand analysis methods assume that each driver behaves the same way on weekdays. We try to confirm this by analyzing panel data.����������
Choice of Departure Time�
In general, drivers are aware of their arrival time at destination because of arrival constraint such as
work start time.�
【～90% of expressway users use ETC system!!】
Pooled data is obtained by putting together all vehicle data over the observation period. If the same vehicle is observed again it will be treated as a d i fferent vehic le. Conventional methods use pooled data and assume that driver behavior is stable over time.�
Panel data, on the other hand, recognizes the same vehicle on another observation. Thus, we can check if there is variation is driver behavior.�
There are 5 data points and �5 different vehicles!�
There is 1 vehicle with 5 data points!�
Day of week analysis
Month of year
Results show that number of trips each driver makes is not stable for each day of the week. In general, number of trips made by drivers are more regular on Mondays. �
Lower level of %WPSS means driver makes more regular number of trips.�
%W
PS
S
%W
PS
S
In general, drivers tend to reduce their regularity during the end of the year.�
BBeehhaavviioorr iiss nnoott ssttaabbllee oovveerr ttiimmee !!��
Departure time
Drivers are aware of being late than b e i n g e a r l y o n M o n d a y s a n d Wednesdays. On Tuesday, Thursday and Friday, late and early penalty are slightly different.�
DDoo ddrriivveerrss aallwwaayyss ddeeppaarrtt eeaarrllyy//llaattee?? We set a Reference Start Time (RST) to evaluate if driver is late or early.
Buffer time = 95th Percentile travel time for a trip – Average travel time (Lomax,et. al. 2003)
ETCデータを用いた都市高速の交通需要変動分析�
* This research was conducted under the collaborative study with Tokyo Metropolitan Expressway Co. Ltd. �
