I think this has biggest advantage in military submarine communications.

You launch a drone underwater. Drone flies off, sends radio message, records another from headquarters. Then flies to preprogrammed place. Then dives to another preprogrammed place and is recovered by sub.

This would me major improvement over communication buoys. Currently when buoy is transmitting, you can guess that the sub is 0,5km away in the general direction of movement of the buoy.

Subs are not really weight limited, so you could have a 20+ mile long cable attached to a buoy giving 2 way communication. Which creates ~1200+ square mile search area assuming you can't just follow the cable.

But they are volume limited, though I believe they generally store the towed array sonar in the ballast tanks, so the same thing could be done.

More serious problems are the time it would take to extend and retract the buoy, speed limitations while the buoy is being deployed, and noise mitgation on the buoy and cable.

Right now, submarines carry disposable communication buoys that can be deployed from depth and sent a pre-recorded message via satellite. Submarines can be alerted to a pending message while at depth through Extremely-Low-Frequency radio communication that can penetrate water, at which point the submarine can move to periscope depth and communicate via satellite without actually surfacing.

Subs are weight limited. There is current trend to make control surfaces out of glass fiber to save weight. The vessel has to be buoyant despite 80mm thick steel skin. Spain just had trouble with too heavy sub. http://www.dailymail.co.uk/news/article-2336953/Spains-1-75b...

Long cables are strength limited. At some point you have to put floats here and there so it doesn't break under it's own weight

And then there is the parasitic drag. Long cables have large surface area.

Neat stuff. Still needs tethering, which makes sense while it's being developed.

However, reality is no match for imagination. When I was a kid the TV show Voyage to the Bottom of the Sea was popular. One of the absolute coolest things was the Flying Sub[1][2]:

   36 foot wide and long, flying submersible,
   aptly called the "Flying Sub"
   ...
   It was deployed through bomb-bay like doors.
   As it broke the surface, its engines could
   generate enough thrust for the vehicle to
   take off and fly at supersonic speeds.

[1] https://en.wikipedia.org/wiki/USOS_Seaview#Refit_and_the_Fly... [2] https://www.youtube.com/watch?v=WHNNXVZ1mYg

This is cool but the autonomy problem remains. Most (aerial) drones can't fly more than 15 minutes. I am curious about the power used underwater to move it.

The ~15 minute lifetime comes from needing to continuously tread air to stay aloft. (You can use up the battery faster by going fast, but there is a max time to just hover.) In contrast, a water-density vehicle could float around and leisurely collect data for almost unlimited amounts of time so long as it doesn't need to get anywhere quickly. (Currents will complicate this.)

15 minutes is a long time. My Trex 450 in idle up mode has about 4 to 5 minutes max of flight time.

Reminds me of one of my all-time favorite Wikipedia articles: https://en.wikipedia.org/wiki/Tradeoffs_for_locomotion_in_ai...

If the public can come up with them surely there are some cool military versions already out there. While it is interesting what was shown in the video it clearly is not designed for two mediums. If anything you would add features like putting the motors on gimbals and feathering of the motors not used. You should not need to tilt the whole thing to change directions underwater, surely a better solution exist

This isn't the public. This is military-funded research at Rutgers. The military generally doesn't have a reason to fund things it has (although DARPA might just not know about classified technologies developed by different branches). But in most cases, the military funds things in academia, a fraction of which become militarized later.

Now we "just" need proper batteries...

Sounds great, if you stop reading before the second-to-last sentence:

"Because radio transmission through water is difficult, the craft must be tethered at this time to provide continuous communications."

Artificial intelligence will help with that. For high value activities like detecting underwater mines and collecting data for underwater oil spills, etc., there should be enough funding to make these autonomous.

That's not the fundamental problem if you want a real-time communication link for something like, say, an inspection or a search and rescue operation - transmitting radio through water is. Water is a pretty bad medium for transmitting RF energy.

You could use ultrasound.

Looks neat, but underwater, it seems more than a big awkward in maneuvering. Makes me curious if a tiltrotor design would be better here in terms of efficiency and speed.

The possibilities here are quite diverse.

They say underseas drones are the future.

Sadly (horrifyingly?) they are also going to be used for nuclear weapons.

Horrifyingly, drones are already being used as nuclear weapons. The Russians recently 'leaked' a new torpedo design with all the characteristics of a drone, being able to find its way through thousands of miles of ocean, finding a target and if this target is acceptable destroying a whole harbor zone. Doomsday drones --or machines if you will-- are likely already in place and able to decide to blow up the world in an instant. I don't like thinking about the capability of these superpower nations.

Also submarine launched radio controlled "drones" carrying nukes have been around since 1955.

https://en.wikipedia.org/wiki/SSM-N-8_Regulus

The only reason why it's called "cruise missile" is because it's self destructing. That's hardly of any tactical importance when talking about nukes.

> Doomsday drones --or machines if you will-- are likely already in place and able to decide to blow up the world in an instant.

What do you base that on?

No facts, only that it's relatively easy to implement such a system. Add to this the fear that the other party may have implemented a doomsday device, and it becomes easy to rationalize the need for such a device.

I would even argue the operators in nuclear bunkers are already part of a doomsday device if they have orders to automatically launch an attack if their country has been struck by nuclear weapons.