Skip to Content

How to Make UAVs Fully Autonomous

A smarter vision system could help robotic aircraft detect airborne obstacles.

Unmanned aerial vehicles (UAVs for short) have proved their usefulness as military tools. But most UAVs aren’t truly autonomous: they’re operated remotely by a human controller from the ground.

To become truly autonomous, UAVs will need to get far better at sensing obstacles and reacting in time to avoid a collision. This will be especially important if they are ever to operate in commercial space.

Sanjiv Singh, a professor and researcher at Carnegie Mellon University, has developed a new system to help UAVs do just this.

Since most UAVs are fairly small and lightweight, they can’t carry the heavy, power-hungry sensors that larger aircraft can use to detect other planes. So Singh and student Debadeepta Dey developed an algorithm that uses an ordinary camera and several software programs to detect potential obstacles.

Their sense-and-avoid system functions across a wide field of view (from up to three miles away) and in a wide range of weather conditions. It does this by finding contrasting points in a video image (such as a dark spot against white clouds) and tracking them to determine movement.

In the video below, the system outlines moving objects in red, such as a plane (distinguished by the green box). It also identifies the characteristic movement of dust–rather than a flying obstacle–on the lens (blue).

Click here to see a bigger version of the video.

“We have proved that sense and avoid for unmanned aerial vehicles using passive sensors is a very real possibility, and with some more time and maturity, this will evolve into a deployable standard technology,” says Dey, who presented details of the system at the International Conference on Field and Service Robotics yesterday.

The sense-and-avoid system can pick out a small, two-seater plane from five miles away, says Dey. So far, he and Singh have tested it from the ground using real aircraft. Currently, it produces some false positives (identifying bugs as planes, for example), but the researchers plan to couple a lidar sensor to the camera to improve it. By bouncing a laser beam off of the obstacle, the lidar will measure its distance to help determine whether it’s really a plane on a collision course or just an insect hitching a ride.

Keep Reading

Most Popular

light and shadow on floor
light and shadow on floor

How Facebook and Google fund global misinformation

The tech giants are paying millions of dollars to the operators of clickbait pages, bankrolling the deterioration of information ecosystems around the world.

protein structures
protein structures

DeepMind says it will release the structure of every protein known to science

The company has already used its protein-folding AI, AlphaFold, to generate structures for the human proteome, as well as yeast, fruit flies, mice, and more.

ASML machine
ASML machine

Inside the machine that saved Moore’s Law

The Dutch firm ASML spent $9 billion and 17 years developing a way to keep making denser computer chips.

brain map
brain map

This is what happens when you see the face of someone you love

The moment we recognize someone, a lot happens all at once. We aren’t aware of any of it.

Stay connected

Illustration by Rose WongIllustration by Rose Wong

Get the latest updates from
MIT Technology Review

Discover special offers, top stories, upcoming events, and more.

Thank you for submitting your email!

Explore more newsletters

It looks like something went wrong.

We’re having trouble saving your preferences. Try refreshing this page and updating them one more time. If you continue to get this message, reach out to us at customer-service@technologyreview.com with a list of newsletters you’d like to receive.