DYNAMIC LATE MERGE SYSTEM EVALUATION: INITIAL DEPLOYMENT ON U.S. ROUTE 10 SUMMER 2003

The Minnesota Department of Transportation (Mn/DOT) is assessing new traffic control strategies for lane closures in work zones. The traditional work zone performs adequately as long as the traffic demand does not exceed the capacity of the single lane closure. When congestion occurs, dangerous travel characteristics develop due to incompatible driver behaviors and presumptions about correct merging procedures. There are two types of merging behavior observed in advance of highway work zones. The first and most common is an early merge where drivers leave the closed lane in advance of the congestion queue. The second merge type is a late merge where drivers use all lanes to the physical merge point, this type is only observed where drivers are instructed to do so, or where drivers are uninformed, such as in incident areas. Unconventional traffic control strategies have been used in attempt to control these two merge types. These systems have shown benefits over the standard work zone traffic control under certain conditions, however, none of these systems respond to these changing conditions. The objective of this project was to develop, test, and evaluate a traffic control system that dynamically incorporates the best aspects of both the Early and Late Merge systems. The Dynamic Late Merge System (DLMS) is a fully automated system that utilizes Doppler radar and Remote Traffic Microwave Sensors (RTMS) to collect data on the current state of traffic. Under predefined traffic conditions, Changeable Message Signs (CMS) display lane-use instructions to drivers in the section of roadway preceding the lane closure. The deployment of the DLMS shortened queue lengths by 35 percent, equalized lane usage near the taper point, but slightly reduced vehicle volume through the construction zone. The instructions of the Dynamic Late Merge System were seldom followed exactly forming multiple merge locations resulting in unintended stop and go conditions. During the lane closure, the number of vehicles exiting at one-half mile upstream of the taper location increased approximately 40 percent indicating that many were willing to use other routes to reach their final destination. Implementation of the DLMS was very brief. During this time, a noticeable change was observed in driver behavior. Drivers were beginning to utilize the lanes in accordance with the instructions on the CMS as the construction period progressed. Unfortunately, drivers never continually approached the merge point to take turns merging. This first deployment in a series did yield some useful knowledge about sign placement and data collection for use in future operations of the DLMS. This initial deployment was the first step in assessing the overall effectiveness of a Dynamic Late Merge System.