Nationwide surveys of departments of transportation, metropolitan planning organizations, and classification vendors/producers were conducted to determine the state of practice on equipment and methodologies used to determine truck vehicle miles traveled (VMT). The current Illinois Department of Transportation (IDOT) methodology was evaluated and it was found that it overestimated truck VMT for multi-unit trucks on all eight functional classes except on the minor urban arterials. The average overestimation was 11.5% and it varied from -10% to +44%. The current method overestimated truck VMT for single-unit trucks in five and underestimated in three functional classes. The under/over estimation ranged from -6% to +35%, but the average value was close to zero. To calculate truck VMT more accurately, this study proposed two different methods based on average truck percentage (ATP) and average section length (ASL). In the ATP method, truck VMT is calculated by multiplying the ATP for a group of roadway sections by the total VMT of that group. The ATP method should be used when the ATP and the total VMT by volume groups are available. In the ASL method, the total truck volume for the sampled sections is multiplied by the ASL. The ASL method should be used when the information required for ATP is not available or not reliable. Sample size influences the accuracy of truck VMT estimation and the decision on sample size must consider the error level that is acceptable. This study looked at the likely error for different sample sizes and recommended using 8% to 16% of the number of roadway sections. The sections should be distributed among the volume groups. Recently, IDOT collects vehicle classification data for three categories at about 10,000 sections, biennially. It is recommended to evaluate the truck VMT calculation using recent data.

Vehicle Miles Traveled (VMT) is a critical measure of highway system performance used extensively in highway transportation management not only for reporting to oversight agencies such as the FHWA but also as an input for financial analysis, resource allocation, and impact assessments. In the current era as highway revenue from fuel taxes continues to fall and direct user charging such as VMT fees become increasingly attractive, consistent and reliable VMT estimates have become critical for evaluating highway funding options. In the current practice at most highway agencies including the Indiana DOT, there exists several alternative methods for VMT estimation that typically yield a spectrum of estimates that are inconsistent and for certain methods, even inaccurate. This study was commissioned by the Indiana Department of Transportation (INDOT) to develop a benchmark method for VMT estimation and to provide calibration factors for adjusting the VMT estimates derived from the other VMT estimation methods. The benchmark method used in this study was a segment-level framework that decomposes the entire road inventory into links and for each link, determining the product of the traffic volume and the inventory length. For the state highway system, the entire population was used; a comprehensive database was developed which facilitates extensive aggregations of VMT by geographical scope, route, functional class, and vehicle class. For the local roads, a sample of counties of different spatial locations and degrees of urbanization were used, and cluster analysis, geographic information systems (GIS), and spatial interpolation techniques were used to expand the VMT estimates from the local road samples to the population of all counties in the state. The results of this study indicate that there is significant variation, with respect to the benchmark method, of the VMT estimates of the other estimation methods. An implementation platform was developed in this study to produce outcomes that address the VMT data needs of the intended end users and stakeholders; this can be expanded to include new roads in future. It was determined that the current statewide VMT (2013) is 78 billion vehicle-miles, which is expected to grow to 95 billion vehicle miles in 2035.

A variety of transportation related applications such as safety, geometric and pavement design of roadways depend on reliable estimates of heavy truck VMT. Truck VMT estimation methodologies used by state DOTs fall under two broad groups; the non-traffic count based and the traffic count based methods. The latter is the most preferred among state DOTs as it utilizes actual data of vehicle movement on a road segment. Resource constraints, however, make it impactical for the monitoring of all road sections of interest continuously throughout the year. State DOTs therefore maintain a traffic count program comprising of permanent counts where data are collected all year round and short-term counts usually collected for periods up to 48 hours. The short-term counts do not represent an average annual daily count. Conversion of the short-term count data to annual average daily estimates are achieved by applying adjustment factors developed from the permanent count data to account for temporal variations. It has been observed that trucks do not exhibit the same temporal variation patterns as passenger vehicles, however, the current practice by a majority of state DOTs involve the use of adjustment factors derived from aggregate volume data not truck volume which fail to adequately explain the temporal variations in truck traffic resulting in biased annual average daily truck traffic (AADTT) estimates. This research utilizes Automatic Traffic Recorder (ATR) data for the rural primary roadways in Iowa to compare three AADTT estimation methods; truck adjustment factor, annual truck percentge and count specific truck percentage methods. An assessment of the acuracies of the 3 methods is made using the estimates of prediction error obtained from cross-validation. Pairwise comparison fo the methods is done by using the non-parametric bootstrap statistical analysis.