AUTONEWS

Real-time driver warning system helps prevent animal collisions

It's 9 p.m. on a rural mountain highway when a 750-pound elk steps onto the road, its silhouette hidden from an oncoming driver by a curve in the terrain.

Suddenly, a roadside sign illuminates: "Wildlife Detected Ahead. Slow Down." The driver eases off the accelerator, and both travelers continue on safely.

This scenario became a reality along a 14-mile stretch of Route 460 in Southwest Virginia that recently served as a one-year pilot project testing a real-time wildlife detection and automated driver warning system. Led by Virginia Tech Transportation Institute (VTTI), the system used digital road signs to alert motorists to the presence of elk and other large animals. Initial results showed drivers responded to the warnings by reducing their speeds by an average of 3 mph when approaching activated signs.

“We have demonstrated an automated system that can detect wildlife, activate real-time warnings, and produce a measurable change in driver behavior both at the warning sign and across the broader detection corridor,” said Shreyas Bhat, a research associate at VTTI.

Sponsored by the Virginia Department of Transportation (VDOT), the system utilizes a network of thermal cameras and AI software to detect medium- to large-sized animals in the roadway and trigger a live warning on roadside signs for approaching drivers. The VTTI team partnered with Crosstek, a provider of off-the-shelf cameras and AI software for wildlife detection, to develop the system and leveraged T-Mobile's 5G cellular network and satellite service for remote monitoring.

Throughout deployment, the team collected data on more than 10,000 passing vehicles. The detection system itself demonstrated an overall accuracy of 97 percent for identifying medium- and large-sized wildlife.

The project’s innovative approach to remote data management recently won first place in the 2026 T-Mobile Innovate Awards Innovation in Community category.

The pilot was part of VDOT’s expansion of Corridor Q, which extends Route 460 from the Kentucky state line to the town of Grundy, Virginia. Associated with the Appalachian Development Highway System, Corridor Q traverses portions of Kentucky, Virginia, and West Virginia, providing an additional interstate connection to those in rural Appalachia. The corridor is also home to approximately 100 elk. 

Collisions involving elk are costly, averaging more than $80,000 per crash, according to the project’s feasibility study. With an adult male elk weighing between 550 and 1,300 pounds, the risk to drivers is high. At the same time, elk tourism in Buchanan County has become a popular economic driver for the area.  

“The motivation was to reduce the risk of accidents or fatalities while also protecting the elk in their habitat,” said Jean Paul Talledo Vilela, a VTTI research scientist.

The challenge...While similar systems have been deployed on other roadways in the U.S., building a solution that could operate reliably in remote, mountainous terrain brought its own challenges.

Talledo Vilela, the lead engineer on the project, said system architecture had to be tailored toward the unique demands of operating in the Appalachian Mountains.

“This system is unique because of the requirements of the area — it needed to be solar powered, portable, and be able to withstand all weather conditions,” Talledo Vilela said.

The wildlife detection and warning system consisted of eight portable, solar-powered trailers and dynamic message signs deployed on a 750-meter section of the Route 460 corridor.

VTTI collaborated with Crosstek to integrate the trailer network with thermal cameras that continuously monitored the roadway and surrounding shoulder areas for elk and other large wildlife. When these animals were detected, the roadside warning signs activated. 

“The topography made communication signals challenging,” Talledo Vilela said. “We needed to ensure that all the trailers had the ability to communicate with each other and send the data from each of those back to the main hub.”

As part of the effort to measure driver behavior, roadside radar units were deployed along the corridor to analyze vehicle speed.

“Speed is an important data point to understand if drivers were actually paying attention to the message board,” Talledo Vilela said. “Even if they are not braking aggressively, some reduction in speed indicates that they may be paying more attention to their surroundings because they got that alert.”

Looking ahead...The initial pilot ended in September 2026. Although data analysis is continuing, Talledo Vilela believes the project demonstrates how portable wildlife detection systems could be deployed in other areas where animals frequently cross roadways.

“One of the goals was to understand how a portable solution can be transferred or adopted in other locations,” Talledo Vilela said.

Beyond wildlife detection, the research team sees opportunities to adapt the technology for other roadway safety applications that require real-time hazard detection and driver warnings.

Roadkill hot spot...It's the first time the technology has been used in New South Wales, with the trial being conducted on a 300 metre section of the Monaro Highway at Michelago, between Canberra and Cooma.

"This is an area that has been assessed as having one of the highest rates of animal wildlife strikes on the network in New South Wales,"  Ms Aitchinson said.

Wildcare Queanbeyan president Elena Guarracino described the area is "an incredible hot spot" for kangaroos in particular.

"There are a couple of dams on one side of the Monaro Highway, and they're continually crossing there and it's just a very … horrible area," she said.

"I've counted up to 30 dead kangaroos at that spot. Along that road towards Williamsdale every 10 metres you'll see a dead kangaroo.

"That's partly fuelled by the ski traffic, but also there's a lot of traffic just generally that commutes to Canberra along that road.

"What we see from the road is just a fraction of what's been hit and moved away from the road until it succumbs to its injuries."

An estimated 10 million animals are killed on Australian roads each year at a huge cost to motorists, insurers and native animals — as well as to wildlife volunteers. 

"When those animals are struck, the wildlife group has to send people out to rescue that injured animal or potentially euthanise it," Ms Guarracino said.

"And then there's the impact of the joey that's in the pouch which, when it's rescued, can take up to two years to rehabilitate before it can be old enough to be released.

"That costs at least $1,200 and enormous volunteer hours — that's if everything goes well and vets aren't involved."

'Smart' version of traditional sign...Ms Guarracino said research had shown the traditional yellow and black wildlife warning signs along roads and highways "don't work".

"They might initially warn the motorists that there is wildlife in that area, but drivers get habituated to those signs," she said.

The digital sign being used as part of the trial, on the other hand, will only display words when activated to alert drivers. 

"What this is about is targeting the message so that it's exactly correct for what is happening on the network," Ms Aitchinson said.

"So it just means that drivers will get more of a warning that will be more appropriate to actually whether there's animals there or not."

The development of PAWS and the subsequent trial is a result of recommendations from the 2024 Wildlife Symposium in Sydney, exploring how emerging technology could be used to reduce wildlife-vehicle collisions.

The AI model behind PAWS has been developed by Transport for NSW's technology and biodiversity teams, combined with animal data from the Symbio Wildlife Park near Wollongong.

Specialist cameras filmed how different species moved and behaved in a controlled setting to build an image library, which was then cross-referenced with time-of-day and weather data to sharpen the system's accuracy in different types of conditions.

"This is an investment of around half a million dollars in just developing the technology," Ms Aitchison said.

"And the benefit of it being done and developed in-house with Transport for New South Wales is that we can ensure that we have the technology, we can use it."

Passengers urged to help lookout for wildlife...The trial at Michelago is due to begin next month, and Transport for NSW will track whether there is a reduction in wildlife strikes in the test area.

Ms Guarracino is hopeful that if the technology proves successful it will not only be rolled out to other hot spots, but also expanded to bring real-time alerts into cars as they drive past.

"It'd be great if the cameras could be integrated into guideposts and give a warning to a driver [via Bluetooth] through their warning devices in their cars," she said.

In the meantime, she continues to urge drivers and passengers to keep their eyes peeled for animal activity, especially during dusk and dawn. 

"I drive very carefully and scan the sides of the road the whole time, even in the daytime," she said.

"And instead of having passengers watching their phone, they need to be scanning the road and helping the driver and going, 'Oh, kangaroo on the left, slow down'."

Real-time animal-collision warning systems...Advanced systems based on artificial intelligence (AI) and video/thermal monitoring are transforming highway safety, helping to prevent accidents between vehicles and wild animals without obstructing wildlife migration routes.

Key technologies and recent projects:

• VTTI and Route 460 (USA/Virginia): A pilot project completed in September 2026 by the Virginia Tech Transportation Institute (VTTI) utilized eight solar-powered, portable trailers equipped with thermal cameras and radar along a stretch of Route 460 to detect elk. The system achieved 97% accuracy and triggered dynamic message signs to alert drivers.

• LAARMA (Australia): Developed in northern Queensland to protect cassowaries, the LAARMA (Large Animal Activated Roadside Monitoring and Alert) system employs low-cost roadside units with AI sensors that activate flashing illuminated signs when the approach of an animal is detected.

• Transport for NSW (Australia): In 2026, this agency tested the PAWS system in areas with high activity of kangaroos, deer, and wombats, combining thermal/optical cameras and computer vision to activate smart signage along the roadway.

• Connected Vehicles (Volvo): New models (such as the XC90 line) incorporate cooperative safety technologies and "Large Animal Alert" features, utilizing automatic braking and steering assistance when the driver's reaction time is insufficient.

• Light/Laser-Based Systems (Netherlands): Trials are exploring the use of thin lines of light or laser along the roadside—triggered by the presence of animals—designed not to block wildlife, but to provide a visual signal to drivers, prompting them to reduce their speed. 

• Initiatives in Brazil: Research led by local institutions, with support from organizations such as FAPESP, is developing computer vision models trained to identify vulnerable species (such as giant anteaters, wolves, and tapirs) and send real-time alerts to mobile apps or vehicle dashboard displays.

Virginia Tech Transportation Institute in Southwest Virginia-USA