Professor Coifman receives prestigious NSF CAREER award to study traffic congestion on freeways.

Traffic congestion on freeways: using probe vehicle data to understand bottlenecks and mitigate the resulting problems. Freeway traffic congestion impacts the movement of most persons and goods in the United States. This congestion is due to a small number of points on the network where demand peaks or capacity drops. Although these bottlenecks typically occupy a short distance of roadway, the resulting queues can extend for several miles. Most freeway queues are characterized by slow and stop waves propagating upstream from a bottleneck. The changing speeds give rise to an increased probability of accidents and the frequent accelerations increase vehicle emissions. Although the propagation of these waves can be predicted once they form, it is not known how they originate and debate continues as to whether they impact capacity. To date, these issues have been studied with point detectors, which are capable of monitoring traffic at fixed locations.

Professor Coifman's sponsored research under his NSF CAREER award will equip two probe vehicles with GPS receivers and distance sensors to monitor adjacent vehicles. The probe vehicles will collect data throughout bottleneck regions and observe important signals that may not propagate all the way to point detectors. These probes will be used to identify which factors influence bottleneck capacity, and identify how the signals and waves form in freeway queues. These efforts should lead to additional benefits including improved car following and traffic flow models. The work is interdisciplinary, drawing on electrical engineering for sensor technologies, and civil engineering for traffic flow theory. Professor Coifman's work also forms the cornerstone for a new course, cross-listed between the two academic departments and taught with a teamwork environment that is found rarely in coursework. In this integrated environment, the students will push the boundaries of knowledge by the end of each quarter.