, Former Professor, Civil, Environmental and Geo-Engineering
This project aimed to improve traffic signal operations from an integrated corridor perspective. In particular, it dealt with a frequently occurring scenario in traffic operations: a freeway gets congested due to an accident, and dynamic message signs suggest drivers use a diversion route, oftentimes a parallel arterial road. If not managed properly, traffic signals on the parallel arterial could become over-saturated from the large volume of diverting freeway traffic. Incidents on arterial networks, including vehicle crashes, severe weather conditions, etc., can also create over-saturated traffic conditions characterized by residual queue and queue spillover to the upstream intersection.
The Integrated Corridor Management (ICM) approach has drawn increased attention in recent years as a promising tool to mitigate urban traffic congestion. In this project, a maximum-flow-based control model was first developed to handle oversaturated traffic conditions at signalized arterials. Based on the arterial control model, an integrated control model was proposed to manage network congestion. Through diversion control, the model aims to fully utilize the available capacity along parallel routes. The impact of the diversion traffic is considered, especially for signalized arterials, so that traffic congestion on the diversion route can be reduced or eliminated by proper adjustment of signal timings. This model does not rely on time-dependent traffic demand as model inputs and is ready to be implemented at typical parallel traffic corridors where the standard detection system is available. The performance of the proposed model was tested using microscopic traffic
simulation in the I-394 and TH 55 corridor in Minneapolis, Minnesota. The results indicate that the proposed model can significantly reduce network congestion.
- Project number: 2011039
- Start date: 02/2011
- Project status: Completed
- Research area: Transportation Safety and Traffic Flow
Congestion, Data and modeling