Understanding Travel Behavior and Vehicle Emissions from GPS and Diary Data: An Application to Southern California.

摘要

The purpose of this thesis is to explore the impact of socio-economic characteristics of drivers on travel behavior and on vehicular emissions of various air pollutants using microscopic data. My starting dataset was collected by SCAG in 2001 and 2002 during their post 2000 Census Regional Travel Survey. Of the 16,939 households who answered the survey, 297 provided self-reported 24-hour travel diary data and detailed GPS data for their vehicles, which was instrumented for SCAG's survey. After selecting 100 out of these 297 households based on their socio-economic characteristics and the completeness of their answers, I relied on 2003 imagery in Google Earth to match diary and GPS data. An extensive clean-up of this dataset yielded a sample of 701 trips, for which I estimated emissions of CO, CO2, NOx, HC, PM10 , and PM2.5 using OpMode in EPA's MOVES2010 (Motor Vehicle Emissions Simulator) from second-by-second GPS travel data. A statistical analysis of the results reveals that men make longer trips than women, although the difference in their emission rates is not statistically significant. Moreover, people 60 or older are the greenest drivers: their driving patterns are more environmentally benign because they accelerate/decelerate less than younger people. Finally, I found significant differences in emission rates based on different household income levels.