Despite their relatively small population, heavy-duty diesel vehicles (HDDVs) are (in 2005) disproportionate contributors to the emissions inventory for oxides of nitrogen (NOx) and particulate matter (PM) due to their high individual vehicle emissions rates, lack of engine aftertreatment, and high vehicle miles traveled. Beginning in the early 1990s, heavy-duty engine manufacturers began equipping their engines with electronic sensors and controls and on-board electronic computer modules (ECMs) to manage these systems. These ECMs can collect and store both periodic and lifetime engine operation data for a variety of engine and vehicle parameters including engine speed and load, time at idle, average vehicle speed, etc. The University of California, Riverside (UCR), under a contract with the California Air Resources Board (CARB), performed data analysis of 270 ECM data sets obtained from the CARB. The results from this analysis have provided insights into engine/vehicle operation that have not been obtained from previous on-board datalogger studies since those previous studies focused on vehicle operation and did not collect engine operating data. Results indicate that HDDVs spend a considerable amount of time at high-speed cruise and at idle and that a smaller percentage of time is spent under transient engine/vehicle operation. These results are consistent with other HDDV activity studies, and provide further proof of the validity of assumptions in CARB’s emission factor (EMFAC2002) model. An additional important contribution of this paper is that the evaluation of vehicle ECM data provides several advantages over traditional global positioning system (GPS) and datalogger studies: (1) ECM data is significantly cheaper than the traditional method ($50 record vs. $2000 record) and (2) ECM data covers vehicle operation over the entire life of the vehicle, whereas traditional surveys cover only short periods of surveillance (days, weeks, or months). It is worthwhile to note that this work was not intended to compare the various methods of data collection but to provide additional empirical support for the EMFAC2002 model and to explore the utility of this unique low-cost form of data collection and analysis. r 2005 Elsevier Ltd. All rights reserved.