Characterizing the effects of driver variability on real-world vehicle emissions

Recent studies on real-world automobile emissions measurements have not adequately addressed the question of whether driving style affects emission levels. In this study, we hypothesized that given the same experimental conditions and a random selection of drivers, the variability associated with individual driving styles (e.g. intensity or duration of acceleration events) would produce statistically significant differences in measured emissions. To test this driver variability hypothesis, we conducted a field study on 24 drivers in a single vehicle on a specified route under low traffic conditions using on-board exhaust emission and engine operating data analyzers and tested for statistically significant differences in CO and NOx emissions between drivers. Our data show significant (95% level) variations in carbon monoxide (GO) and oxides of nitrogen (NOx) emissions among the 24 drivers under driving conditions where we have controlled for driving route, traffic density and vehicle type. Since the ANOVA tests showed significant differences in emissions between drivers but the frequency of driving modes were very similar, this suggests that the intensity of vehicle operation within a give mode, not the modal frequency, explains the emissions variability between drivers. (C) 1998 Elsevier Science Ltd. All rights reserved.