New app warns drivers of lane departures

Silver car driving around a curve in a rural paved road

Unintentional lane departures are a leading factor in most fatal crashes that occur on Minnesota’s highways. Lane-departure crashes typically involve human behavior—for example, a tired long-haul fleet driver drifting off to sleep and off the road. But a new smartphone app developed by U of M researchers may help drivers stay in their lane.

The app, developed in research funded by the Minnesota Local Road Research Board (LRRB) and the Minnesota Department of Transportation (MnDOT), was the subject of a November 14 LRRB webinar.

Victor Lund, traffic engineer for St. Louis County and technical liaison for the project, opened the webinar by noting that 60 percent of fatal crashes in Minnesota involve driver inattentiveness, drowsiness, or impairment leading to roadway departure. “Given the opportunity, drivers will put themselves in bad positions,” Lund said. 

Lund explained that the geometric design of our state’s roads adds regional risk factors: 80 percent of all fatal crashes occur on state and county highway systems and 60 percent of those fatal crashes occur in rural Minnesota. Southern Minnesota’s grid-based network consists primarily of tangent, or straight roads, while in Northern Minnesota, a curvilinear network navigates around rivers and lakes. Those curved sections—45,000 miles of roads across the state—are where half of roadway departure crashes occur.

“That emphasizes the need on the traffic safety infrastructure side to focus on curves,” Lund said. “But we’re also looking at technologies to help drivers be more alert and give them warnings so that when they are potentially departing a lane, they can take corrective action.”

It turns out, those technologies already exist. A research team led by CTS scholar Imran Hayee, professor and chair of the electrical engineering department at the U of M Duluth, used GPS technology to power an app that monitors a vehicle’s position and the upcoming road geometry to calculate risks in real time. If a vehicle begins to stray from its lane, an alert sound prompts the driver to steer back. (See a demonstration.) 

Highly connected, sensor-rich new vehicles may include built-in lane-departure technologies that use camera or vision-based systems. However, poor weather, obscured lane markings, and individual vehicle factors can impair reliability. When a camera cannot detect lane markings or a sensor is coated with snow, a driver must rely on fallible human senses. Hayee also noted that rumble strips—an effective physical road engineering solution—are only on installed on the right side of the road, while a vehicle may depart its lane to the right or left. The LRRB project addresses those possibilities with an algorithm that uses a standard GPS receiver interacting with a cloud-based backend browser tool and Google routes.

“With some processing, we can estimate the lateral shift on the road and issue a warning if there is a shift without a vehicle indicator,” Hayee said. “We added a connectivity feature so that any vehicle can travel on a particular road and upload its trajectory to our cloud, where the algorithm can extract the road reference from that vehicle to be available for use by any other vehicle.”

This technology currently exists as an Android app, but at this point requires an external GPS device to operate. Future implementations could include an iOS app, integration with Google and Apple Maps, and in-vehicle integration. “It’s especially useful for long stretches of rural roadways where trucks are driving at nighttime,” Hayee said—although all drivers can benefit from a digital eye on the road.

—Amy Goetzman, contributing writer

Subscribe

Sign up to receive our Catalyst newsletter in your inbox twice every month.

Media Contact

Michael McCarthy
612-624-3645