Global Positioning System-based Truck Modelling for Regional Travel Demand Forecasting

16 Freight movement complexity and a lack of truck trip data limits freight modelling’s practicality 17 and application. Most publicly available truck movement data is reported at higher geographical levels 18 than detailed analysis and modelling require. The current Freight Analysis Framework (FAF) provides 19 freight data based on the Commodity Flow Survey (CFS) conducted every five years with commercial 20 vehicle data collected from state departments of transportation (DOT) and other proprietary data sources. 21 Commercially available commodity flow databases (e.g., Global Insight’s TRANSEARCH) are 22 excessively expensive, and this data set also provides aggregate information on commodity shipments 23 between selected major cities. However, this data set provides limited information on shipments into 24 smaller geographies. Freight transportation surveys such as roadside/intercept surveys, focus and 25 stakeholder group surveys, and commodity flow surveys are time-consuming and requires a great amount 26 of labor to geocode the trip origins and destinations. 27 This study shows a truck modelling effort utilizing real-time global positioning system (GPS) 28 truck data with a belief that GPS-based databases may provide detailed and disaggregate data for regional 29 freight modelling. Eight weeks of GPS data for 5,000 different trucks in 2011 (two weeks in four 30 seasons) for the Atlanta and Birmingham metropolitan planning organizations (MPO) was collected and 31 incorporated with other datasets. Notably, GPS data can provide detailed origin-destination (O-D) 32 information, routes and corridors, operating speeds, travel times, and flows. This study explores various 33 possible ways that a GPS-based truck movement data can contribute to freight demand forecasts at the 34 state and regional levels. 35