A variable speed limit system used to regulate traffic flow through work zones on a 7.5 mile section of I-495 saved motorists approximately 267 vehicle-hours of delay each day.

Experience with variable speed limits on I-495 in Virginia

Date Posted

Work Zone Variable Speed Limit Systems: Effectiveness and System Design Issues

Summary Information

In July 2008, a variable speed limit (VSL) system was implemented on a 7.5 mile section of I-495 (the Capital Beltway) in Virginia between the Springfield Interchange and the Woodrow Wilson Bridge. A study was conducted to examine performance and evaluate system impacts, and a literature search was performed to identify measures of effectiveness (MOEs) and gather information on VSL system and algorithm design.
A simulation model (VISSIM) was used to evaluate system impacts based on traffic volume data provided by the VDOT Traffic Monitoring System. Several MOEs were evaluated including travel time, speed, queue length, number of stops, and lane change information. Travel time and mean speed served as operational measures, and the standard deviation of speed, mean queue length, and mean number of lane changes (per 5 minute interval) served as surrogate measures of safety. The model was calibrated using field data collected from floating car runs made between June and September 2008. Researchers identified where vehicle speeds dropped as a result of work zones and made adjustments to the model to match real-world travel times within 10 percent.

Overall, the system implemented on the Capital Beltway was designed to harmonize upstream and downstream traffic flow around each work zone. The work zones were configured with tapered lane closures that reduced travel lanes from 4 lanes to 2 lanes or from 4 lanes to 1 lane. In addition to calculating optimal speeds, the VSL system was configured to limit the frequency of speed limit changes. Previous research suggested that 2-minute intervals between posted speed limit changes would be detrimental to safety, whereas 5- or 10-minute intervals would reduce crash potential (Lee, et al., 2004).


Benefits were computed by examining differences in travel time between the base case with static speed limits and the best performing VSL alternative. The simulation model indicated that the system can improve traffic conditions if the demand does not exceed capacity by too large a margin. For example, in the case of the 4-to-1 lane closure, none of the VSL alternatives evaluated produced an increase in network speeds. However, in the case of the 4-to-2 lane closure, the best VSL configuration resulted in a mean savings of 267.04 vehicle-hours of delay, which was translated into $12,229.76 user delay savings per day in the control area when the lane closure was present.

With an assumed system cost of $3.2 million, researchers estimated that the system would pay for itself (in terms of user delay savings) after 262 days of operations. Overall, findings indicate that these systems are best suited for long-term deployments and not for short-term temporary deployments.
Goal Areas
Deployment Locations