Mihir Garimella is making drones that go where humans cannot

Editor's Note: (This feature is part of Tomorrow's Hero, a series profiling young innovators transforming the world for a brighter future. Discover their stories here.)

(CNN) When Mihir Garimella was in fifth grade, his parents would chide him for leaving his bedroom light on. Rather than just remember to turn off the light, he rigged a motion sensor to automate the switch.

A year later at his school orchestra, Garimella would get in trouble because he couldn't tune his violin. So he built a robot that listened to sound samples and tuned it for him.

The 18-year-old from Pittsburgh, Pennsylvania is an ordinary teenager with an extraordinary eye for integrating tech solutions into everyday life. A Google Science Fair winner in 2014, a TEDx speaker in 2015, he'd designed a low-cost concussion detection kit, created a math app which recognizes and solves handwritten problems, and co-authored a paper on brain tumor diagnosis via image processing algorithms -- all before leaving high school.

"I've always loved just building things," he told CNN. It's perhaps an understatement.

High-tech solutions from nature

Garimella is best known for his work on drones, for which he won the Google prize. His research has taken him into the field of biomimicry -- designing robotics modeled on the natural world. In Garimella's case, he's focused on the humble fruit fly, an unlikely source of inspiration.

"Four years ago my family went to India on vacation," he recalls. "When we got back we realized we'd left a bunch of bananas on the kitchen counter. By that point they were rotting, so our house was filled with fruit flies."

Try as he might, they'd always evade the fly swat. "I became curious about how these tiny organisms (with a) tiny brain, really bad eyesight, could possibly escape so effectively," he says.

Garimella studied the insect, and sent emails to eminent biologists. To his surprise, professors at Caltech and Carnegie Mellon University replied.

"Fruit flies have the fastest visual system on the planet -- they can see 10 times faster than humans can," Garimella says. "Their eyes are equivalent to a 25x25 pixel camera, so they basically can't see any detail at all. But as a result, they can process whatever information they do see very quickly."

The teenager began wondering if this low-data approach could be applied to drones, traditionally difficult to maneuver in tight spaces and ill-adept in detecting and avoiding danger.

His tiny FlyBot prototype fitted in the palm of a hand and acted like a fruit fly, outmaneuvering hazards using a single camera and Garimella's own algorithm to 3D-map its environment. The prizes flowed in, but for the student, there was more work to be done.

Going where humans dare not

"The next thing I looked at was carrying out missions in these spaces -- search and rescue, finding the source of a fire or chemical leak," he says. Finding positive uses for drones has been a hot topic in recent years, spurring on stories about medical drops in Rwanda and Amazon deliveries and competitions like Drones For Good in the UAE.

Often these drones take different forms, varying in size. Garimella has been working to design a single drone that can perform multiple functions.

"The idea is that you take (my) drone, plug in different sensors based on whatever task you're trying to accomplish, and the drone would use those sensors to carry out a mission," he says. "So in a search and rescue situation after an earthquake, you could plug in a thermal camera and the drone would use that to find people who are trapped.

"If you're inspecting a nuclear power plant you could plug in radiation sensors and the drone could use that to pinpoint hot spots of radiation, find tanks that are going to leak."

He lists many more potential uses, including industrial inspection. The idea is that with an adaptive, autonomous drone, these robots can spot potential dangers before they ever happen.

"I'm trying to keep these drones under the $1,000 price range," he adds, saying that after four years of work, he hopes to have a product on the market in the next "one-to-two years."

As a freshman studying computer science at Stanford University he's now balancing studies while launching his own company.

"My dream is to build something that can improve the lives of a billion people," he says.