A simulation study in Minneapolis-St. Paul estimated that ramp metering decreased total system travel time by 6 to16 percent and increased average mainline speeds by 13 to 26 percent.
Date Posted
06/13/2002
Identifier
2007-B00482
TwitterLinkedInFacebook

Evaluation of Ramp Meter Control Effectiveness in Two Twin Cities Freeways

Summary Information

This study used simulation modeling to evaluate the performance of various ramp meter control strategies implemented in Minneapolis-St. Paul. The model was developed as an alternative to ramp meter shut-off testing as a way to measure performance and optimize ramp meter timing without disrupting traffic.

The model was built around the AIMSUN (Advanced Interactive Microscopic Simulator for Urban and Non-Urban Networks) model, and was designed to emulate traffic conditions based on data collected from traffic detectors in the Twin Cities. In addition, the simulation was able to modify ramp meter timings and compensate for congested conditions in real-time by predicting driver behavior and estimating impacts of traveler information on traffic conditions.

The integrated ramp meter control strategy was designed to handle freeway zones with unidirectional links extending 3 to 6 miles in length. The zone of influence for each ramp included the upstream free-flow area (e.g. low incident area) and the downstream bottleneck area (e.g., lane drop areas, high volume entrance ramps, and weaving areas). The control algorithm was designed to balance the volume of traffic entering and leaving each zone every 30 seconds. When the density of vehicles decreased, additional capacity was recognized.

In order to develop a model that would accurately reflect the geographical setting at each ramp, a graphical-editor for traffic networks was used to create an individual zone of influence for each ramp meter. The graphical editor accepted spatial data from computer drafting programs or aerial photographs and facilitated the placement of nodes in a model network. Once the ramp meter network was integrated into AIMSUN, traffic detector data was loaded as model input. The input consisted of field data (5-minute volume and occupancy measurements) collected from 3000 TMC Mn/DOT detectors, and manual traffic counts determined from cameras and on-site observations. Conditions at model boundary areas were determined through manual counts at TH-169 and I-94. All data represented incident free conditions between 2:00 PM to 8:00 PM.

SIMULATION RESULTS
  • Ramp metering decreased total system travel time by 6 to16 percent.
  • Ramp metering increased average mainline speeds by 13 to 26 percent
  • The number of stops was relatively low on the mainline with active ramp metering; however, the total number of stops increased by 1000 percent without ramp metering
  • Overall, wait times were tolerable at ramp meters; however, in a few cases wait times measured 21 minutes. The author noted these extreme cases were similar to the actual wait times experience by some Twin Cities commuters prior to the ramp meter shut down
Goal Areas
Deployment Locations