Validating the Performance of the FHWA Work Zone Model Version 1.0: A Case Study Along I-91 in Springfield, Massachusetts

Berthaume, Andrew; Jackson, Lauren; Berg, Ian; O'Donnell, Brian · 2017 · ROSA P / John A. Volpe National Transportation Systems Center (U.S.)

archive: archived pipeline: cataloged verified

Get this paper ↗ (full text — opens at the source; we link to it, we don't host it)

Summary

This paper validates the performance of the Federal Highway Administration (FHWA) Work Zone Driver Model Version 1.0, a software tool designed to improve the accuracy of traffic microsimulation in freeway work zones. The research addresses the limitation of existing microscopic simulation models, such as VISSIM’s default Weidemann 99 car-following algorithm, which were not specifically calibrated for the unique driver behaviors observed in work zones. As infrastructure aging increases work zone activity and associated delays, accurate prediction tools are critical for developing effective Traffic Management Plans. The FHWA model, implemented as a dynamic link library (DLL), overrides standard car-following logic in commercial software to simulate driver responses based on psycho-physical frameworks and Modified Field Theory derived from doctoral research. The study employs a case study approach, interfacing the FHWA DLL with VISSIM 7.0 and testing it against field data collected from an instrumented research vehicle (IRV) on Interstate 91 in Springfield, Massachusetts. The test site was a 6-mile segment featuring a long-term work zone that reduced capacity from three lanes to two. Network calibration utilized MassDOT traffic volume and speed data. The FHWA model’s performance was compared to both the field observations and the native Weidemann 99 model across three key metrics aligned with state Department of Transportation interests: travel time through work zone segments, travel speeds, and back-of-queue locations. Simulations were run ten times for each model, while field data comprised seven IRV runs during the AM peak hour. Results indicated that the FHWA DLL performed acceptably, often outperforming the Weidemann 99 model in specific areas. The FHWA model more accurately predicted average travel speeds across all segments and instantaneous speeds at critical points. Crucially, the FHWA model correctly identified queue locations, predicting queues forming in the taper zone as observed in the field, whereas the Weidemann model incorrectly placed queues in the advanced warning zone. However, the FHWA model was less accurate in predicting total travel time compared to Weidemann. Both models failed to reproduce the variance in travel times, speeds, and queue lengths observed in the field data. This lack of variance is attributed to the FHWA DLL’s current limitation: it only overrides car-following logic and lacks algorithms for other behaviors such as lane changing, conflict zone yielding, and route choice, which contribute to traffic shock waves and variability. The study concludes that while the FHWA Work Zone Driver Model v1.0 provides improved accuracy for specific work zone metrics like speed and queue location, it requires refinement. The authors recommend that Version 2.0 include acceleration and deceleration algorithms for non-car-following behaviors, such as gap acceptance and cooperative merging, to better capture traffic variance. Additionally, they suggest that mixing FHWA and Weidemann vehicles in simulations could optimize results by leveraging the strengths of both models. These findings support the continued development of specialized microsimulation tools to help practitioners better predict and mitigate work zone impacts.

Key finding

The FHWA Work Zone Driver Model DLL v1.0 predicted queue locations and travel speeds more accurately than the standard Wiedemann 99 model, though both models failed to reproduce the variance in field-observed travel speeds and queue lengths.

Methodology

simulator

Provenance

The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed. Discovered via bulk_ingest_rosap on 2026-05-23 (6 acquisition events logged).

Summary generated by qwen3.6-27b-prismaquant on 2026-06-10; verification: verified.

Topics

Ranked by relevance to this paper. Hover a topic for its definition.

• work zones

Information type

What kind of knowledge this paper contributes, grouped by family — independent of topic (what it is about) and method (how it was studied).

• Theoretical Contribution: computational model