Statistical Approach to Estimating Truck Traffic Speed and Its Application to Emission Inventory Modeling

Abstract

Heavy-duty diesel trucks are a significant source of carbon dioxide, oxides of nitrogen, and particulate matter emissions. The construction of accurate emission inventories of these trucks requires a proper characterization of their speed because emission rates vary substantially by speed. However, data regarding truck speed have been very limited to date. This paper presents a statistical method for estimating truck traffic speed that takes advantage of the existing traffic monitoring systems. With traffic data from these systems examined, it was found that truck traffic speed can be effectively estimated on the basis of the knowledge of the overall traffic speed alone. The relationships between the two variables are shown to be very strong, with the coefficient of determination of the regression equations being .98 or higher. All regression coefficients are found to be statistically significant at the 5% alpha level. In addition, the validation results of the estimated truck traffic speed show a good estimation performance of the regression equations, where most of the average absolute errors are within the range of 2 to 4 mph. On the basis of the developed statistical relationship, regional truck activity, in terms of vehicle miles traveled versus speed distribution, on Southern California freeways was estimated and used to construct the associated running exhaust emission inventories. The resulting emission inventories show that using the heavy-duty diesel truck-specific speed distribution rather than the overall speed distribution reduces the estimates of oxides of nitrogen emissions by 4% and particulate matter emissions by 26%.

Fusion of Vehicle Weight and Activity Data for Improved Vehicle Emission Modeling

“…On Wednesday, April 15, 2009, there were 12 trucks estimated to pass by VDS 718479 during the 5-min period from 00:40:00 to 00:45:00. The overall traffic speed at this VDS was 66.2 mph and the estimated truck traffic speed was × = 66.2 mph 0.863 57.1 mph where 0.863 is the coefficient of the linear relationship between truck traffic speed and overall traffic speed derived in a previous study (3). Since the effective length of this VDS is 0.33 mi, the temporal footprint of these trucks at this VDS during this 5-min period was…”

“…Several efforts have been made by researchers to combine data from different sources to create better HDT activity data input for MOVES. For example, previous research developed statistical models based on truck traffic speed from a WIM station and overall traffic speed from a nearby VDS, so that truck traffic speed at other VDS can be estimated based on the knowledge of the overall traffic speed alone (3). The research presented in this paper is an extension of the previous research.…”

“…In addition, PeMS VDS only report speed estimates for the overall traffic without differentiating the trucks. Previous analysis of truck speed data from 15 WIM stations in southern California and overall traffic speed data from their closest PeMS VDS found that there is a strong linear relationship between truck traffic speed and overall traffic speed, although the relationship may vary from site to site (3). In this research, truck traffic speed at each PeMS VDS was estimated based on the results from that previous research.…”

A Review on Machine Learning for Sustainable Construction Equipment

“…Frey et al [21] explored onboard mass emission data and developed model using case studies for multiple construction equipment. Regression analysis technique was used by [22,23] respectively to quantify the impact of idling and estimating tra c speed of equipment to optimize the emission. Using ML and inertial sensors, [24] developed a model to predict emission level of construction equipment.…”