Highly responsive sense-and-avoid drone technology is essential if we’re ever going to see the likes of Amazon using autonomous flying machines to deliver packages to customers.
And which hobbyist quadcopter pilot wouldn’t want to see effective avoidance capabilities incorporated into their own drone, a feature that would surely lead to fewer wallet-busting repair bills following clumsy flight maneuvers.
If the work of Ross Allen and Marco Pavone is anything to go by, we’re closer than ever to the rollout of advanced versions of existing avoidance technology. The pair, both researchers at Stanford University’s Department of Aeronautics and Astronautics, are working on a leading-edge sense-and-avoid system that’s smart enough to dodge the pokes and strokes of a fast-moving fencer.
The video above shows the technology in action, and we’re sure you’ll agree, it’s pretty incredible. This isn’t a case of a drone having time to see an object like a building or pylon before taking action to move out of the way. It’s responding super-fast to a moving object – in this case a fencer’s sword – and darting out of the way to avoid contact.
Allen and Pavone’s impressive system, once perfected, looks as if it’d work comfortably indoors with many stationary obstacles in a relatively tight flying space, and also outside in a more unpredictable environment that could potentially present a more diverse set of challenges.
At the current time, the technology pulls data from an external camera before processing it and transmitting it to the drone, but the researchers are working on incorporating directly into the copter a range of visual, laser, and ultrasonic sensors for the same functionality, Popular Science reported.
Avoidance technology is already starting to find its way into some consumer drones. DJI’s recently released Phantom 4 machine, for example, uses multiple on-board cameras for its “obstacle sensing system” that can react to objects between 0.7 and 15 meters away.
It’s certainly pretty cool to see the sword-dodging drone in action, though it might not do much to steady the nerves of those living in fear of a robot uprising. Hopefully Allen and Pavone won’t ever equip this particular flying machine with its autonomous drone technology, or else we’ll definitely be in trouble.
- The DroNet algorithm teaches drones to navigate city streets like cars
- MIT drones navigate more effectively in crowded spaces by embracing uncertainty
- During the Drone Rodeo at CES, the fastest pilots race for aerial supremacy
- Amazon’s delivery drones could hitch rides on trucks to save power
- Take to the skies with the best drones under $500