a drone built by NVIDIA researchers navigates the most far-flung, un-mapped places using deep learning and computer vision powered by NVIDIA ‘jetson TX1’ AI supercomputers. initially designed to follow forest trails to rescue lost hikers or spot fallen trees, the low-flying autonomous drone could work in canyons, between skyscrapers, or inside buildings where GPS is restricted, or unavailable.


all images © NDIVIA 

 

 

to keep costs low, the researchers built their device using an off-the-shelf drone equipped with the NVIDIAjetson TX1‘ and two cameras.

‘our whole idea is to use cameras to understand and navigate the environment,’ says nikolai smolyanskiy, the team’s technical lead. ‘jetson gives us the computing power to do advanced AI onboard the drone, which is a requirement for operating in remote environments.’

 

 

the team has already trained it to follow train tracks, and ported the system to a robot-on-wheels to traverse hallways. the drone also avoids obstacles like people, pets or poles. although the technology is still experimental, it could eventually search for survivors in damaged buildings, check stock on store shelves, inspect railroad tracks in tunnels, or be adapted to examine communications cables underwater.

 

 

the autonomous drone learned to find its way by watching video that smolyanskiy shot along eight miles of trails in the pacific northwest of america. he captured the video in different lighting conditions using three wide-angle gopro cameras mounted on the left, center and right of a metal bar on a mini segway.

 

 

the team now plans to create downloadable software for ‘jetson TX1’ and ‘jetson TX2’ AI supercomputers so designers can build robots that navigate on visual information alone. the idea is to command the robot to travel between two points on any map — whether it’s a google map or a building plan — and have it successfully make the trip, avoiding obstacles along the way.

Save

    have something to add? share your thoughts in our comments section below.

    comments policy
    LOG IN
    designboom's comment policy guidelines
    generally speaking, if we publish something, it's because we're genuinely interested in the subject. we hope you'll share this interest and if you know even more about it, please share! our goal in the discussion threads is to have good conversation and we prefer constructive opinions. we and our readers have fun with entertaining ones. designboom welcomes alerts about typos, incorrect names, and the like.
    the correction is at the discretion of the post editor and may not happen immediately.

    what if you disagree with what we or another commenter has to say?
    let's hear it! but please understand that offensive, inappropriate, or just plain annoying comments may be deleted or shortened.

    - please do not make racist, sexist, anti-semitic, homophobic or otherwise offensive comments.
    - please don't personally insult the writers or your fellow commenters.
    - please avoid using offensive words, replacing a few letters with asterisks is not a valid workaround.
    - please don't include your website or e-mail address in your comments for the purpose of self-promotion.
    - please respect jury verdicts and do not discuss offensively on the competition results
    (there is only one fist prize, and designboom usually asks renown professionals to help us to promote talent.
    in addition to the awarded designs, we do feel that almost all deserve our attention, that is why we publish
    the best 100-200 entries too.)

    a link is allowed in comments as long as they add value in the form of information, images, humor, etc. (links to the front page of your personal blog or website are not okay). unwelcome links (to commercial products or services of others, offensive material etc. ) will be redacted. and, ... yes, spam gets banned. no, we do not post fake comments.

    technology news