Groundbreaking AI tech aims to reduce wildlife collisions — and it was invented by teen girls

Two images. On the left, a deer crosses a road, with oncoming traffic in the distance. On the right, four teen girls smile in front of a row of Samsung Galaxy tablets

Every year in the United States, there are between one and two million collisions between cars and large animals, according to the Federal Highway Administration.

In Colorado in particular, the state’s Department of Transportation reports about 4,000 wildlife collisions annually, which costs an estimated $80 million for Colorado drivers every year. While wildlife crossings and other interventions have helped, this issue still puts the lives of people and animals at risk.

So, a group of teens at STEM School Highlands Ranch, in a suburb just outside of Denver, decided to do something about it.

These teenagers — a group of four young women — have created a wildlife detection device using infrared cameras and artificial intelligence. They call it “Project Deer.”

Four teen girls sit in a classroom, holding up a certificate for their project. Their male teacher kneels in front of them, smiling
Photo courtesy of STEM School Highlands Ranch

“The current plan is to utilize an infrared camera to image the surroundings in front to detect animals and warn the driver when animals are in the environment so they can pay more attention, slow down, and be much more likely to avoid a collision,” Bri Scoville, one of the student inventors, told CBS Colorado.

Infrared cameras detect body heat (even in compromising conditions like darkness or inclement weather). From there, an algorithm tracks the heat and motion, and AI machine learning categorizes the data and logs it as an animal.

“The use of infrared is definitely new and something that hasn’t been done before,” added Siddhi Singh, a junior at STEM School Highlands Ranch.

Once the AI model detects wildlife, the detection device sends a signal to a smaller device inside of a vehicle to warn drivers with a combination of light and sound, allowing them to slow down or stop. 

An infrared sensor is connected to a computer
The students used infrared sensors to create their device. Photo courtesy of STEM School Highlands Ranch

The team — which also includes fellow students Dhriti Sinha and Robyn Ballheim — assembled for of Samsung’s annual Solve for Tomorrow competition.

“Generally, what Samsung asks you to do is to find a problem in your community that's affecting your community and then try to use technology to solve the problem,” Tylor Chacon, the team’s sponsor and computer science teacher, told CBS Colorado. 

The girls won the state contest for Colorado, which came with a prize of $12,000 to help continue building their prototype. (They also received a number of Samsung Galaxy Chromebooks.)

Fortunately, the device isn’t as costly as other complex inventions; each detection device is made of four $5 infrared sensors. Creating something accessible and ubiquitous is vital in confronting an issue as widespread as this threat to wildlife.

While the invention did not win awards at the national level, the team is continuing to perfect their design, with the help of experts in the automotive industry, as well as Colorado’s DOT.

Although CDOT has stationary wildlife sensors, the STEM School team hopes their device would be more effective — and easier to take on the go.

“We spoke with someone at Audi, they said that there are huge hopes for this device if we’re able to get this to work,” Singh told CBS Colorado.

“Getting it to work,” is more of a matter of when; not if.

Over summer break, the teens are trying out different approaches to make the prototype the best version it can be, including the use of Bluetooth systems, or mounting the device like a dashboard camera on a vehicle’s windshield.

A video still showing a mockup of the inside of a car alerting a driver to oncoming wildlife. It reads: "The driver is alerted with audio and imagery."
Photo courtesy of STEM School Highlands Ranch

“That’s our ultimate goal, to engineer something that anyone and everyone can put on their cars,” Singh said.

The tech is there. In fact, the team tested the prototype on their school’s therapy dog and saw resounding success.

“We were able to implement a combo of heat tracking, motion tracking and classification to build the best detection system,” Singh told Denver7 this spring. 

“As a proof of concept, we strapped all the components of our design on a remote control car and with permission, tested it on Daisy, our school's therapy dog.”

Now, it’s just a matter of creating a final product that can do this vital signaling in those life-or-death moments. The team told CBS that they plan to re-enter the Samsung solve for Tomorrow competition next year, with a refined prototype.

A chocolate lab in a pink harness sits on a lawn next to a pile of tennis balls
Daisy the therapy dog is often included in student projects. Photo courtesy of STEM School Highlands Ranch

“There’s really no limit to how far the project could go,” Chacon added. “I think it’s within sight to fix this problem.”

Winning is certainly the goal, but with ongoing national coverage and support from major industry players, they also want to keep other young women in STEM from feeling like a deer in the headlights.

“We… hope that we can encourage other girls to innovate like us,” Scoville told CBS.

The team’s mentor, Chacon, sees a world of success for this project — and for those other young girls may dream up themselves.

He said: “These girls are laying the groundwork for somebody to look up to them later on and say, ‘I wanna do that, I wanna be that, look what they did, I can do that, too.’”

Header images courtesy of Can Pac Swire (CC BY-NC 2.0) and STEM School Highlands Ranch

Article Details

June 24, 2024 12:37 PM
A woman using a tablet to read braille.

An 8th grader invented an accessible braille reader that could help people save thousands of dollars

1.3 million people in the US have impaired vision, but only 10% read braille. A new, inexpensive braille reader could change lives.
A person wearing flippers drags a seal-shaped decoy across the water. The decoy is lit by streaks of LED lights, visible from its underbelly.

LED surfboards could light the way for fewer shark attacks, scientists say

Researchers at Macquarie University have discovered a way to trick sharks into leaving surfers alone.
No items found.

Want to stay up-to-date on positive news?

The best email in your inbox.
Filled with the day’s best good news.