April 27, 2011
According to the National Highway Traffic Safety Administration, in the past ten years there have been no advances in technology to reduce traffic accidents involving pedestrians. To address this major issue for New Jersey's urban areas, a Senior Design Team at Stevens Institute of Technology is working with the New Jersey Department of Transportation (NJ DOT) to develop a novel traffic signal system that protects pedestrians without interrupting traffic flow.
The project and the team was developed by Stevens alumnus Andy Kaplan '09, a Traffic Engineer with the Transportation Safety Resource Center at Rutgers University. Members of the Senior Design project come from three engineering backgrounds: Abel Alvarez, Mechanical Engineering; Kyle Brisson, Electrical Engineering; Eric Chirlin, Computer Engineering; Cassidy DeSchryver, Mechanical Engineering; and Jeffrey Lichtenfeld, Computer Engineering. The team's faculty advisors are Dr. Michael Zavlanos from Mechanical Engineering and Dr. Hong Man of Electrical and Computer Engineering (ECE).
"This is a perfect example of the sort of interdisciplinary project the Stevens undergraduates are uniquely prepared for," says Distinguished Service Professor Bruce McNair, the ECE Senior Design course coordinator. "It is the sort of project that enables students to connect their studies with real-world needs."
After meeting with NJ DOT engineers to gather statistics and other data, such as the typical length of a red light and how long it takes the average individual to cross the street, the team identified weaknesses in the current traffic signal system. Specifically, they saw that pedestrians do not feel that the current system works for them. The "walk" button lacks feedback, prompting impatient pedestrians to cross against the light, and the static nature of traffic signal timing does not respond to changes in pedestrian and automobile traffic.
The team's solution, Ped-Aware, is a traffic signal and detection system that provides two advances in traffic systems. First, by using traffic signals, their system passively detects pedestrians and provides feedback as to when the light will change. Second, Ped-Aware uses both pedestrian and vehicle data to switch traffic signals dynamically.
These two components allow Ped-Aware to change the right-of-way based on the actual status of an intersection, rather than rely on pre-programmed signal cycles. By making intersections more efficient, the system promotes pedestrian safety without compromising the performance of vehicular traffic.
In order to detect pedestrians, Ped-Aware uses cameras that compare images over time to deduce when and how many pedestrians are waiting to cross a street. Using MATLab, the team developed image processing software based on existing algorithms. Image data is processed in real time, without a recording unit, so there are no privacy concerns.
To test their idea, the team created a program that simulates a traffic intersection running Ped-Aware. Simulation results demonstrate that the system effectively gives both pedestrians and cars enough right-of-way to avoid congestion.
NJ DOT has already shown interest in Ped-Aware's scalable and adaptable dynamic traffic signal system. After they deliver their final presentation at Stevens, the team hopes to also demonstrate their solution to urban traffic problems to NJ DOT engineers.