Drones’ Challenge: Add ‘Autonomy’ in Flying

MADISON, Wis.—Most consumer drones today are flying blind.

“It’s too easy to crash a drone,” said Philip McNamara, a serial entrepreneur and organizer of Drones Data X Conference. Too often, consumers send their drones—which can cost more than $1,000—slamming into a wall or soaring into the grasp of a tree.

Adding “autonomy” to flying drones – a pre-requisites for the proliferation of drones – would involve “a large body of work” that has barely begun, McNamara explained.

(Source: DJI)

(Source: DJI)

Taking on the challenge of autonomous flying are DJI, the world’s leading maker of drones and aerial cameras based in Shenzhen, China, and Movidius (San Mateo, Calif.), a designer of low-power computer vision SoCs.

Remi El-Ouazzane, Movidius CEO, told EE Times, “Efforts to achieve full autonomy for drones have been accelerating.” Whether drones are used for package deliveries or building inspection, they require a collision avoidance feature—critical in autonomous flying, he said.

One of the ways to add autonomy is to enable drones to see.

Based on a Movidius vision-processing SoC called Myriad 2 VPU, DJI launched its Phantom 4 aircraft, with “the ability to sense and avoid obstacles in real time and hover in a fixed position without the need for a GPS signal,” according to Movidius. That’s “an industry first in making advanced visual guidance systems a standard feature for consumer drones,” the company added.

The DJI Phantom 4 is able to sense objects and respond by avoiding, or halting flight until further instruction (Source: DJI)

The DJI Phantom 4 is able to sense objects and respond by avoiding, or halting flight until further instruction (Source: DJI)

Drones can now sense an object and fly around it, dodge people who get in their way, and go from point A to B. That’s quite a feat. But drones have reached only “the first level” of full autonomous flying, in El-Ouazzane’s opinion. “They can do a lot more in the future,” he added.

By integrating computer vision that can learn on the fly, drones will be able to do “scene labeling, semantic extraction and ultimately a lot more neural networks-based learning,” he explained. Drones, for example, can identify a specific person and serve as his flying videographer.

Why DJI?
McNamara called DJI “the absolute leaders in the drone space.” He told EE Times that other companies are only “scampering behind them, trying to nip at their heels.”

To read the rest of this article, visit EBN sister site EE Times.

1 comment on “Drones’ Challenge: Add ‘Autonomy’ in Flying

  1. Luscombe854
    May 6, 2017

    others firms have actually made main expansions so that we can make   WhatsApp  You will certainly get some search results page below the search box


Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.