A Variable Speed Limit System Implemented on a 15-Mile Section of I-95 in Virginia Reduced Fatal and Injury Crash Rates by 22 Percent and Rear-end Crash Rates by Nine Percent.

Virginia Study Evaluated the Effectiveness of the Proactive Variable Speed Limit Algorithm.

Date Posted
02/26/2026
Identifier
2026-B02027

Interstate 95 Variable Speed Limit System

Summary Information

I-95 is a critical north-south corridor along the eastern seaboard that carries heavy passenger and freight traffic. The northbound segment of I-95 between Exit 126 (US 1/Route 17) and Exit 143 (Route 610) in Virginia was ranked as the seventh worst traffic bottleneck in the nation in 2017. To address these recurring safety and operational issues, this study evaluated the performance of a Variable Speed Limit System (VSL) on this corridor, activated in June 2022.

METHODOLOGY

The project focused on the development, testing, and deployment of a proactive VSL system designed to harmonize traffic flow and reduce crash risk under congested conditions. Detector data were used as primary inputs to a custom congestion-prediction algorithm that dynamically determines appropriate speed limits within a predefined range of 35 to 70 MPH. The algorithm was structured as a three-step process that first identified the prevailing traffic state using real-time detector measurements, then generated site-specific speed recommendations, and finally applied spatial and temporal smoothing constraints to limit abrupt speed changes both over time and between adjacent signs. This design enabled the system to anticipate traffic breakdowns and incrementally reduce speeds in advance of congestion, rather than reacting solely to existing queues. The algorithm was developed iteratively using a microscopic traffic simulation model (VISSIM ) of the corridor. The project team also launched a communications initiative to advertise the VSL system, which was done using physical postings as well as mobile digital advertisements.

FINDINGS

Data from June 22, 2022, to February 20, 2023, showed that rates for all types of crashes were reduced compared with the same period of the year before activation. In specific, fatal and injury crash rates decreased by 22 percent, and rear-end crash rates decreased by nine percent. 

Goal Areas
Results Type