MIT engineers help fight roadside perils

Aim to cut crashes at traffic stops, accident scenes

John DiFava, a former Massachusetts State Police superintendent, is lucky to be alive. In 1976, an out-of-control vehicle nearly killed DiFava as he stood outside his patrol car on Interstate 495. “I was able to jump into the cruiser just in time,’’ said DiFava, who is currently chief of police at Massachusetts Institute of Technology.

Now, 36 years later, about a dozen police officers are killed by careless drivers each year, according to the Federal Bureau of Investigation. Also at risk are other emergency responders, such as firefighters, paramedics, and tow truck operators. And in Massachusetts, there are two or three accidents a month in which a police officer or his vehicle is hit while stopped.

“Some involve just the cruiser being struck with no injuries to the trooper,’’ State Police spokesman David Procopio wrote in an e-mail. “Others have involved injuries,’’ including a trooper whose wrist was hurt Thursday after his cruiser was struck by a pickup truck on Interstate 93 in Milton.


There have been attempts to address the problem, including improved training for responders and new “move over’’ laws requiring drivers to give emergency vehicles a wide berth.

Get Talking Points in your inbox:
An afternoon recap of the day’s most important business news, delivered weekdays.
Thank you for signing up! Sign up for more newsletters here

Now, one of DiFava’s MIT colleagues wants to give technology a try. Professor Seth Teller of the university’s Computer Science and Artificial Intelligence Lab is trying to create life-saving devices for police, paramedics, and other first responders.

Teller is working on two concepts. One is a roof-mounted laser that could project safety warnings onto the road behind the police car. The idea is to eliminate the need to place flares or traffic cones on the roadway, a practice that puts police in danger.

The second is a combined video and radar unit capable of spotting an approaching car and issuing a warning if it threatens to strike the emergency worker or his vehicle.

DiFava is a big fan.


“If they get this equipment to work,’’ he said, “it is going to be almost as much a revolution in safety for policing as soft body armor.’’

Teller‘s project, called Divert and Alert, is aided by a $500,000 grant from the US Justice Department and an alliance with defense contractor BAE Systems Inc. Still, the researcher does not know whether his systems will work, or whether government agencies will be able to afford them. For now, he said, “we just want to see whether we can do this, at any cost.’’

Teller is working closely with the Massachusetts State Police to develop and test prototypes.

“It was exactly the direction we needed to go,’’ said Colonel Marian McGovern, the current State Police superintendent. “Our goal is to keep everybody safe out there.’’

Also working with Teller is an associate professor of aeronautics and astronautics, Missy Cummings, an MIT colleague who was one of the Navy’s first female fighter pilots. Cummings used to fly F/A-18 fighter jets. Today, she’s a “human factors’’ expert, a specialist in simplifying complex technologies. She is working to ensure that passing drivers and parked police officers will each respond correctly to the Divert and Alert systems.


“You’ve got two sets of people who have to be accounted for,’’ Cummings said. “There are a lot of technology issues that have to be worked out.’’

To date, most efforts to protect police from roadside accidents have featured ever-brighter warning lights, Teller said.

“In my judgment, that’s done about as much good as it’s going to do,’’ he said, because the brighter lights can blind approaching drivers, making them even more of a threat. Some scientists believe that humans may actually tend to turn toward bright lights, like a moth attracted to a flame, he added.

Teller’s laser projector would aim its light at the pavement, displaying images that would warn a driver to give a stopped police car a wide berth. The projector might display words like “KEEP LEFT,’’ or it might draw images of traffic cones to visually nudge the driver into the far lane.

“Can we build a device to do this?’’ Teller asked. “We don’t know yet.’’

One question is whether a box mounted on the roof of a police car can throw the projected image far enough down the road. Teller said the projector might have to be mounted on a mast several feet above the car’s roof, which would make the system more cumbersome and costly.

Then there is the weather. The projection system might not work in fog or on snow-covered roads. For now, Teller said he will be content with testing it in clear weather, on a clean road, at night.

“If we can solve that problem, then we’ll move on to rain, snow, fog - all of those,’’ he said.

The video/radar system for spotting approaching cars would not be affected by weather and could pick up vehicles at a great distance. It could sound an alarm, warning of a threat.

But how do you create a system that knows which vehicles pose a threat? Too many cars and trucks zip past in the adjacent lane at full speed, even though “move over’’ laws in most US states require drivers to move one lane over when passing a stopped emergency vehicle, or to slow down when no alternate lane is available.

“The technical challenge,’’ Teller said, “is to distinguish the trajectories of the cars who are doing the nominal thing, and staying in the lane, from the cars that pose a threat.’’

He said he thinks the device could be programmed to spot cars that are swerving or driving well above the speed limit and to look for other hints, such as a car approaching at night with its lights off. “If we see any of those conditions occurring, we can raise the alarm,’’ Teller said.

In addition, Cummings said, the sensor system must sound its warning while the car is far enough away to give a police officer time to respond.

“How early can we detect that a driver is not responding as he should?’’ she said. “When should we actually sound the alarm?’’

Unless the system is consistently correct, it could generate so many false alarms that users would ignore it or switch it off, she said - “the ‘cry wolf’ effect.’’

Teller and Cummings are making no promises.

“That’s why it’s research,’’ Cummings said.

But DiFava said he’s a believer.

“I have developed, in the 10 years I have been at MIT, an incredibly healthy respect for the place,’’ he said. “If anybody can get it to work, they can get it to work.’’

Hiawatha Bray can be reached at