CUAUV Nova Full Mission Test (Downward)

Video from the downward camera of Cornell University's Autonomous Underwater Vehicle (cuauv.org) completing a miniature version of the 2009 AUVSI competition mission. The vehicle autonomously completes all of the visual tasks, including simulating grabbing the briefcase with only visual data.

Category:

Science & Technology

Tags:

• auv
• cornell
• auvsi
• nova
• cuauv

[JosephLisee]

Nice run there, is the video at about 5 times normal speed?

Looks pretty complete except for an actual treasure grab and a marker drop (unless they are too small to see). With all of the swaying your vehicle did, I am impressed it stayed locked onto those bins, good job to your vision folks.

[CornellAUV]

The video is about 2x normal speed, and only every other frame, which makes it look a little fast.

None of the actuators (dropper, grabber, or torpedoes) were on the vehicle. If the correct silhouettes didn't come into frame right away, the vehicle would have aligned with the bins, and the adjustments wouldn't have looked quite so dumb.

[JosephLisee]

So what are you logging frames at, 15 FPS? We have similar computational power to your vehicle and we find logging video sucks up lots of cycles, more then we like actually. So I am curious about recording frame rates, codecs, and bit rates.

[JosephLisee]

Who needs those orange pipes ;).