Print ...

Contributor...

  • Editor:

Path: Top

Route choice observation in Bandung street network.

Master Theses from JBPTUNIKOMPP / 2012-06-30 07:36:31
By : Russ Bona Frazila, S2 - Highway Systems Engineering
Created : 1998-00, with files

Keyword : Urban street network; Route choice behaviour

This thesis describes empirical research into route choice behaviour in an urban street network.

The objectives of the research were to observe route choice behaviour on alternative routes in urban street networks and to model the route choice behaviour according to license plate matching data. The observed network was consisting of two alternative routes, simplified from real road network in the eastern Bandung area. The data was collected using home interviews, with 168 respondents that were making daily trips through the observed network. Also using license plate matching survey (9 hours using video recorder, 39 hours using audio recording). From the home interviews data, it was found that most travelers choose a route by more than one reason, 62.0 of respondents selected three reasons in choosing route and only 4.8 % of respondents considered only one reason for route choice. It was also found that respondents's perception on travel time and journey length through the observed network tended to be overestimated. The license plate

matching survey resulted a very low number of matching vehicles, that was only 8.4 % matched from total recorded license plate on Route 1 and 5.0 % on Route 2. Data from license plate matching survey was used for calibrating some route choice model specifications using only travel times as independent variable. All calibrated models have poor statistical performance. One of the best models from a statistical point of view by its survey methods is as follows: - Using Video License Plate Matching Survey Data : P1 = [1/(1 0.564(t1/t2) super-script 1.019)] R superscript 2 = 0.224; - Using Audio License Plate Matching Survey Data: P1 = 1.26e superscript-0.71 (t1/t2) R superscript 2 = 0.179 Where: P1 = %age of vehicle choosing route 1 (%), t1 = travel time through Route 1 minute), t2 = travel time through Route 2 (minute).

The poor statistical performance of calibrated models could be caused by some factors, such as the limitation of the model that only used one variable (travel time), calibration method, survey methods that did not provide sufficient traveler attributes, less data (survey time).

Description Alternative :

This thesis describes empirical research into route choice behaviour in an urban street network.

The objectives of the research were to observe route choice behaviour on alternative routes in urban street networks and to model the route choice behaviour according to license plate matching data. The observed network was consisting of two alternative routes, simplified from real road network in the eastern Bandung area. The data was collected using home interviews, with 168 respondents that were making daily trips through the observed network. Also using license plate matching survey (9 hours using video recorder, 39 hours using audio recording). From the home interviews data, it was found that most travelers choose a route by more than one reason, 62.0 of respondents selected three reasons in choosing route and only 4.8 % of respondents considered only one reason for route choice. It was also found that respondents's perception on travel time and journey length through the observed network tended to be overestimated. The license plate

matching survey resulted a very low number of matching vehicles, that was only 8.4 % matched from total recorded license plate on Route 1 and 5.0 % on Route 2. Data from license plate matching survey was used for calibrating some route choice model specifications using only travel times as independent variable. All calibrated models have poor statistical performance. One of the best models from a statistical point of view by its survey methods is as follows: - Using Video License Plate Matching Survey Data : P1 = [1/(1 0.564(t1/t2) super-script 1.019)] R superscript 2 = 0.224; - Using Audio License Plate Matching Survey Data: P1 = 1.26e superscript-0.71 (t1/t2) R superscript 2 = 0.179 Where: P1 = %age of vehicle choosing route 1 (%), t1 = travel time through Route 1 minute), t2 = travel time through Route 2 (minute).

The poor statistical performance of calibrated models could be caused by some factors, such as the limitation of the model that only used one variable (travel time), calibration method, survey methods that did not provide sufficient traveler attributes, less data (survey time).


Give Comment ?#(0) | Bookmark

PropertyValue
Publisher IDJBPTUNIKOMPP
OrganizationS
Contact Namedyah@unikom.ac.id
AddressJl. Dipati Ukur No.116 Lt.7
CityBandung
RegionWest Java
CountryIndonesia
Phone022-2533825 ext.112
Fax022-2533754
Administrator E-mailperpus@unikom.ac.id
CKO E-mailperpus@unikom.ac.id