App Battles Driver Distraction but Spares Passengers
A new approach inhibits dangerous phone use by detecting when a driver is on the phone.
Smart phones are a growing source of driver distraction, but researchers hope to use their capabilities to make the devices’ use less dangerous. New sensing technology can determine whether a phone is being used by the driver, or merely a passenger, and is providing a building block for a new generation of distraction-thwarting apps.
Research shows that just talking on a phone increases the risk of a crash four times; texting increases it 23 times. The National Transportation Safety Board (NTSB) recently called for a nationwide ban on driver use of cell phones and other portable electronic devices while driving, and said that electronic devices were a factor behind an unknown number of the more than 3,000 distraction-related deaths on U.S. roads in 2010.
Some existing phone-blocking tools, such as a $20 per year service from iZup, use a smart phone’s GPS receiver to tell when it’s being used inside a moving vehicle, and then automatically send calls to voice mail and delay text messages. Some “driver mode” apps available in Android Market do things like change the interface to feature bigger and fewer buttons on the screen and limit available apps to ones like navigation. And a few wireless carriers offer distraction-fighting services, too.
But these apps have an Achilles’ heel: they are based on detecting car motion. That means that within the car, they can affect not only drivers’ phones but those of passengers, who are present in 38 percent of vehicle trips.
Researchers from Rutgers University and the Stevens Institute of Technology are developing technology that determines when a phone is likely in the driver’s area. It uses the phone to connect with a car’s stereo system via Bluetooth and issue subaudible beeps inside the car. The phone’s microphone picks up the beeps; a signal-processing algorithm calculates the position of a phone within the car.
“We like to find some middle ground—try to reduce the driver distraction caused by phone use, but let the driver decide whether they want to use the phone or not, or use it in a safer way,” says Yingying Chen, a computer scientist at the Stevens Institute of Technology. Chen codeveloped the technology. “We have put all the pieces together so you can get into a car, and when this application runs, it will decide whether you are the driver or the passenger.”
Chen says the technology could be used by existing distraction-prevention apps to make them smarter. But the researchers are also working on a new suite of apps to leverage the driver sensing. One, in development at Rutgers’ Winlab, would notify people on your contact list that you are driving—and may even add prompts asking potential callers whether the call is a true emergency, says Janne Lindqvist, a Winlab researcher developing the app.
Another app in development integrates driver detection with calendars—making it easy to notify participants in an imminent meeting whether you’re running late. “If you are driving, this could become a one-button text message,” says Marco Gruteser, an associate professor at Rutgers. “It’s an interface question on the phone, so you don’t have to go through finding contact information and making the call.”
The Rutgers and Stevens group first demonstrated the driver-detection technology in the lab last year, but have since implemented it on phones and integrated it into apps. They are also working on simplifying the algorithm so that driver detection only takes half of the current seven to eight seconds.
The group plans to conduct studies this spring to determine whether the apps actually change people’s behavior, says Lindqvist. The technology will find its way into commercial offerings within a year, likely through licensing agreements, he predicted.
As well as calling for a ban on mobile device use, the NTSB also called for these sorts of technology solutions. “We have a problem here and it’s growing worse,” says Deborah Hersman, chairwoman of NTSB. “Technology has brought many of these challenges, but technology holds the key to the solution, too. Detecting the driver is a very critical piece.”
The key downside of the Rutgers and Stevens technology is that it relies on Bluetooth, which is absent in most older cars—and not universally available on new models, either. Even in cars with Bluetooth, it is good but not perfect; because of varying cabin sizes and speaker configurations, it accurately detects the driver in 90 percent of cars, and produces false positives in the single percentage range, Gruteser says.
It’s clear that technology solutions are badly needed, says Marcel Just, a psychology professor at Carnegie Mellon University. In a study published in 2008, Just used functional magnetic resonance imaging to show that merely listening to someone speak reduces brain activity associated with driving by 37 percent.
“It seems to me that disallowing cell phone use while driving through some technology fix is a good thing,” he says. “You can imagine it being legislated, and you can imagine insurance companies instituting an undistracted driver discount.” But to be effective, any technology fix would have to be widely implemented, and difficult for the driver to circumvent, he says.
AI is here. Will you lead or follow?
Join us at EmTech Digital 2019.