Most modern vehicles offer some form of lane-keeping system to help drivers stay between the lines when they’re on the road. These advanced driver assistance systems (ADAS) range from basic lane departure warnings to more sophisticated systems that actively correct the vehicle’s course.
These systems rely on predictable pavement markings and expected roadway geometries to function properly. This raises critical questions for transportation agencies: What happens in work zones, where these visual cues are often significantly altered? Could lane-keeping systems malfunction in such environments and elevate the risk of a collision?
The Minnesota Department of Transportation decided to find out. They partnered with SRF Consulting Group and subcontractor ViaSight to conduct a series of high-tech road tests to assess the performance of various lane-keeping systems in work zones — and learn whether work zones could be better designed to accommodate these systems.
“Safety is the number one goal of this project,” said Erik Minge, Project Director in SRF’s Intelligent Transportation Systems/Connected Autonomous Vehicles Group. “The more we can make the roadside infrastructure compatible with the ADAS technologies that are on the vehicles, the better those vehicles can safely navigate work zones.”
The researchers tested four commercially available vehicles with lane-keeping systems, including a Tesla Model Y, a Toyota Corolla, and a Hyundai Elantra. The fourth vehicle, a semi-tractor made by Volvo, allowed the team to gauge the effect of having ADAS cameras mounted much higher off the ground.
Each vehicle was outfitted with cameras and sensors designed to monitor how the lane-keeping system responded to road conditions. Using ViaSight’s AI-powered mobile analytics suite, the team could observe discrepancies between real-world conditions and the ADAS systems’ interpretation of road conditions.
For an initial test location, the team took advantage of MnDOT’s unique pavement research facility, MnROAD, which offers a closed-loop test track that is off-limits to the public. Here, they safely conducted a series of tests on a variety of work zone scenarios involving temporary markings, removed or masked markings, and temporary roadway geometries such as lane shifts.
The team reports that the initial test results are encouraging. All systems performed well in a variety of work zone layouts — only disengaging momentarily in certain conditions. ViaSight’s President and Founder Phil Magney said that while the Tesla’s more advanced sensor suite performed particularly well, with the other vehicles not far behind.
“A couple of the other cars have just intermittent moments where they kind of lose sight of the markings. And that makes sense, because there’s several stretches in here where we have either removed or masked the painted lines,” Magney said.
There’s more work ahead and the team is planning a new series of tests in live work zones later this year to evaluate performance in additional conditions like fresh pavement.
The insights gained from this project will directly help MnDOT choose the best ways to mark lanes in work zones. Clear and more “machine-readable” work zones could mean ADAS systems can perform more reliably. This could lead to fewer sudden disengagements of the system, less driver confusion, and ultimately, a safer environment for everyone navigating these temporary but vital stretches of road.
Gallery
