U.S. Department of Transportation
Federal Highway Administration
1200 New Jersey Avenue, SE
Washington, DC 20590
202-366-4000
Federal Highway Administration Research and Technology
Coordinating, Developing, and Delivering Highway Transportation Innovations
FACT SHEET |
This fact sheet is an archived publication and may contain dated technical, contact, and link information |
Publication Number: FHWA-HRT-18-015 Date: December 2017 |
Publication Number: FHWA-HRT-18-015 Date: December 2017 |
PDF Version (220 KB)
PDF files can be viewed with the Acrobat® Reader®
This fact sheet describes the development of a specialized driver model and software for work zones – the FHWA Work Zone Driver Model v1.1.
The FHWA Work Zone Driver Model v1.1 was created for planners and engineers, so they can more accurately predict the operational impacts of freeway work zones, such as queues, travel times, and delays. The software interfaces with commercially available microsimulation tools and alters car-following algorithms— derived from empirically collected data—to more accurately match behaviors observed in freeway work zones. The software is available for download on the Open Source Application Development Portal (OSADP)1 as part of the FHWA Driver Model Platform v0.6.2
Freeway work zones can have significant operational impacts, accounting for nearly 24 percent of the annual non-recurring delay in the United States.3 With many freeway segments reaching (and surpassing) their 50-year design life, the number of freeway work zones in the United States is projected to increase. So they can design and plan work zones that minimize delays, planners and engineers need accurate simulation tools to predict the operational impacts of work zones.
Microsimulation has the potential to accurately predict work zone impacts, because it models individual vehicle movements at the trajectory level. Existing microsimulation software packages, however, are not equipped to simulate freeway work zones, because their algorithms do not account for changes in car-following behavior (e.g., acceleration patterns) known to exist as drivers approach and traverse freeway work zones. Before practitioners can accurately predict work zone impacts using microsimulation, a microscopic work zone driver behavior model is required.
The model was developed and calibrated using work zone car-following data collected with an instrumented research vehicle (IRV) through freeway work zones on I–95 near Washington, DC in 20134 and on I–91 in Springfield, MA in 2016. Tools were developed to identify and classify car-following instances, filtering and processing data so it could be used to describe car-following for passenger cars traversing freeway segments, advanced warning, taper zones, and work zones. These processed data sets will be available on the Research Data Exchange (RDE).5
After analyzing the processed data, strong variations in car-following behaviors were observed between work zone and non-work zone segments.6 Based on these findings, a car-following model was created that reproduces the unique acceleration patterns, headways, and traveling speeds observed for each segment. The resulting work zone car-following model uses a multi-dimensional psycho-physical framework and acceleration/deceleration algorithms derived from modified field theory to describe car-following for freeways, advanced warning, taper zones, and work zones separately.7,8
Figure 1. Instrumented research vehicle (IRV) used for data collection.4
FHWA and the U.S. Department of Transportation (USDOT) Volpe Center developed a work zone car-following model and DLL file that interfaces with existing microsimulation tools, enabling more accurate simulation of car-following through freeway work zones. The purpose of this project is to provide practitioners at transportation agencies and in private industry with a tool that can better predict the operational impacts of freeway work zones.
Through a technique known as cross-validation, the microscopic models were trained and tested against separate validation data from I–95, ensuring that the acceleration and deceleration behavior were accurate and not over-fit to the empirical data. The software was developed, interfaced with VISSIM, and tested along freeway work zones in Massachusetts.
The Volpe Center worked with MassDOT to test and validate the work zone software against field data collected for the I–91 work zone in Springfield, MA. Performance measures were selected based on how State agencies currently analyze work zone traffic management plans (i.e., location and length of queues and travel times). Field data provided by MassDOT supplement the data collected with the IRV.
The FHWA Work Zone Driver Model predicted travel times and queue locations more accurately than the calibrated microsimulation software package using suggested model parameters. Opportunities to improve the work zone software were noted, including algorithms to reproduce additional variations in driver behavior, currently simulated through microsimulation models’ stochastic elements and random number generators.
Figure 2. Queue locations and lengths for the I-91 work zone, as observed in the field (top), as predicted by VISSIM (middle), and as predicted by the FHWA Work Zone Driver Model (bottom)9
The FHWA Work Zone Driver Model v1.1 is now available as part of the FHWA Driver Model Platform v0.62 (available for download on the OSADP1). The model is being continually upgraded; new versions will be released annually.
For more information, please contact:
Andrew.Berthaume@dot.gov
FHWA-HRT-18-015
HRDO-20/12-2017(200)E