Law-enforcement technologyTackling problem of roadside collisions between drivers, police vehicles

Published 4 April 2012

Police officers, when stopping a speeding car, would park their vehicle on the shoulder of the road, lights flashing, in order to talk with the errand driver; at least one police officer is killed each month in the United States when another driver, out of recklessness, impairment, fatigue, or simple inattention, careens into the stopped police car


Picture this scenario: a police car, lights flashing, is stopped on the side of an Interstate, its occupant helping a motorist with car trouble or writing a speeding ticket. Cars and trucks pass by at high speed.  Then along comes a driver who, out of recklessness, impairment, fatigue, or simple inattention, careens into the stopped police car.

Such scenarios are all too common in the United States. Roadside collisions are one of the leading causes of line-of-duty deaths for law enforcement officers, according to figures from the Federal Bureau of Investigation, with an average of one officer hit and killed by a motor vehicle each month from 2000 to 2009.

Now, a team of researchers at the MIT Computer Science and Artificial Intelligence Laboratory (CSAIL), in collaboration with BAE Systems, has announced a new project to combat this problem.

 An MIT CSAIL release reports that the researchers, working in collaboration with the Massachusetts State Police, will investigate methods to change driver behavior, and warn police when the danger of a collision seems high.

Despite decades of research focused on emergency lighting equipment, improved practices for conducting roadside stops, and perceptual aspects of driving, no one understands the phenomenon of vehicle collisions with stationary police vehicles on the roadside,” said project leader Seth Teller, a professor in MIT’s Department of Electrical Engineering and Computer Science and a principal investigator at CSAIL. “Our goal with this project is to bring together experts in machine vision and human factors to take a fresh look at the problem and develop new technologies to reduce the accident rate and save lives.”

 “Along with other police departments around the country, we have paid a dear price for accepting the risks faced by troopers working on the side of busy highways,” said Colonel Marian J. McGovern, superintendent of the Massachusetts State Police. “Dozens of troopers have been injured in recent years, including the tragic death, in 2010, of Sgt. Douglas Weddleton, who was struck down while protecting a road construction crew. We are eager to work with MIT to explore ways to keep our personnel safe as they perform their duties.”

 Teller observes that while existing vehicle-mounted emergency lighting systems successfully divert some fraction of drivers from veering too close to police vehicles, they fail to have any effect on many others, and they are unable to warn police of seemingly hazardous drivers on the road.

To address these issues, Teller and his research team propose an “end-to-end” approach to the problem involving two mechanisms embedded within a novel piece of equipment mounted on the officer’s cruiser: one to generate dynamic perceptual cues that guide drivers around the stop site, and one to analyze approaching traffic to detect anomalous or hazardous behavior.

Our goal is to develop a robust perception system prototype that can operate at night, in inclement weather, and at sufficiently long distance to alert officers of impending danger, giving them enough time to react while minimizing false alarms,” said Dr. Matthew Antone of BAE Systems.

From an engineering perspective, a substantial challenge is to design this system to be unobtrusive, power efficient, simple to use, and low cost, so that it can be deployed widely and will not interfere with existing equipment or procedures.”

Other team members — including Missy Cummings, an associate professor in MIT’s Department of Aeronautics and Astronautics and a principal investigator at CSAIL, and her colleague Dr. Erin Solovey — will focus on researching the human factors that influence how drivers respond to visual cues and how officers react to an alert system.

MIT researchers expect to have a working prototype of the new system sometime in 2012. The system will be developed and evaluated in collaboration with Massachusetts State Police personnel at the Emergency Vehicle Operator Course, a high-speed training track in Danvers, Massachusetts.

The project is funded by the National Institute of Justice.