The development of a prescreening model to identify failed and gross polluting vehicles

Abstract

The California State Bureau of Automotive Repair uses a high-emitter profile model to direct, or screen a fraction of the vehicle fleet in for inspection and maintenance testing at test-only facilities. Reviews by the California Inspection/Maintenance Review Committee showed the high-emitter profile to be inefficient and in need of improvement. In this study, using in-use vehicle emissions data from California's statewide smog check program, we specified a new multinomial logit model designed to improve the screening efficiency for targeting potential failed and gross polluting vehicles. Modeling results show that factors such as odometer reading, model year, vehicle make, as well as the presence of emissions control systems are significant factors in predicting the likelihood that a screened vehicle will test as a failed or a gross polluting vehicle. Comparisons indicate that the new multinomial logit model specification can predict various inspection/maintenance test outcomes more accurately than the existing high-emitter profile model.

Measuring On-Road Vehicle Emissions with Multiple Instruments Including Remote Sensing

“…Vehicle fleet emissions are dominated by a small percentage of 'high-emitters' with excessive emission levels. This has been confirmed by laboratory test programs [11], remote sensing studies [12] and tunnel studies [13]. This skewness in emission distributions is-at least to some extent-due to the variability in emission profiles by vehicle type, the progressive introduction of cleaner engine, and catalyst technology into the vehicle fleet and ageing effects of the in-use fleet.…”

Investigating the real-world emission characteristics of light-duty gasoline vehicles and their relationship to local socioeconomic conditions in three communities in Los Angeles, California

“…These findings disagree with findings from a study of on-road taxi data in Chicago which found that taxis which accumulate mileage faster than the regular fleet have higher percentage of "high emitters" (Bishop and Stedman, 2008), suggesting a correlation between super-emitters and higher mileage. (Choo et al, 2007).…”

Binary Logistic Regression Modeling of Idle CO Emissions in Order to Estimate Predictors Influences in Old Vehicle Park

“…Unlike binomial logistic regression, in [11] a new multinomial logit model is developed to identify the factors that are significantly associated with identified failed and gross polluting vehicles. The results are similar and also indicate that factors such as odometer reading, model year, and the vehicle make, along with the presence of modern emission control systems, are significant factors in predicting the likelihood being labeled as a failed vehicle and a gross polluter.…”