DARPA's AlphaDog robot pack mule begins real world testing

http://www.extremetech.com/extreme/117493-darpas-alphadog-robot-pack-mule-begins-real-world-testing

by mrsebastian • 14 years ago

53 29 14 years ago

cellis• 14 years ago

But how long before they outfit these things with the nano-quadrotors?

tkahn6• 14 years ago

I always get a serious 'uncanny valley' feeling when I watch a video of one of these things. It reacts very organically when it stumbles and tries to recover and yet the rest of its motions are obvious mechanical facsimiles of natural movement. Reminds me of rat things from Snow Crash.

samstave• 14 years ago

Don't worry, we are right on track for the full cyberpunk reality of 2020.

agumonkey2• 14 years ago

Can't find the video again, there was a humanoid version of this that was frigtheningly humanlike, smooth, articulated, balanced, subtle. It wasnt made to look human, fully opened and wired everywhere, yet the behavior/balance was more complex and on point than anything I've seen.

spyder• 14 years ago

http://www.youtube.com/user/BostonDynamics

capisce• 14 years ago

You're not alone: http://www.youtube.com/watch?v=VXJZVZFRFJc

DasIch• 14 years ago

I'm wondering what would happen if the robot would be attacked? How easy would it be to destroy one? The noise makes it impossible to use for anything that requires a certain amount of stealth. So really what is the use case apart from probably being a very interesting project to work on?

InclinedPlane2• 14 years ago

It's a pack mule. Except pack mules require a lot more infrastructure to take care of and deploy.

If you're in the mountains of, say, Afghanistan, it's really hard to be resupplied. But with something like this you can take hundreds of pounds of food, ammunition, etc. with you into very rugged, unimproved terrain. Which allows you to reduce the load on individual soldiers, increase the length of time deployed soldiers can fight effectively, etc.

WiseWeasel• 14 years ago

With the noise it makes, I'd give this thing unattended about 5 minutes in the mountains of Afghanistan before it gets converted to scrap. Hell, it's such an ear-sore, I might be tempted if I found that thing unattended in the mountains of California...

dsr_• 14 years ago

While voice command recognition is pretty good with a limited command set and a low-noise environment, a battlefield is not a low-noise environment.

OK, anyone could predict that. Let's take it a few steps further:

- military bases are not low-noise environments

- "come here" is a great command, but if there's no authentication, it's a hazard

- in the other direction, just how quiet is the gasoline engine on an AlphaDog? Loud enough to give away your position, relative to the trudge of a squad's worth of boots on the ground?

jessriedel• 14 years ago

The first two problems are pretty much solved by using a microphone headset to issue commands.

schiffern2• 14 years ago

So it costs 32 million dollars, is enormously complicated, has a noisy generator on the back, turns into a brick after 20 miles, and has half the carrying capacity of a camel.

Cool tech, but we've got a ways to go, people.

joe_the_user• 14 years ago

Well, it may indeed be worse than a camel in nearly every respect. But if it's good enough to be widely used, it will also get widely tested, which may allow a next generation item which would be considerably better than a camel.

MostExtremeCake• 14 years ago

Realistic response: a camel requires that someone is always with it, telling it where to go. This thing can just be told to go 5 miles in a direction without supervision and will maybe reliably get there. It can also go through stuff that's too toxic for a camel/human being.

Tin-foil-hat response: Can't strap guns to a camel and have it shoot things. The progression from unmanned tech used for non-combat purposes to combat seems likely. Predator drones used to be used for surveillance and did not have hellfire missiles strapped to them.

abecedarius• 14 years ago

That kinda suggests making an automated cameldriver instead. Of course that's much easier to make fun of.

melling• 14 years ago

Not really, this stuff follows a curve like Moore's Law. Take the first DARPA Grand Challenge for autonomous cars. No one finished, or came even close. A year later, several cars finished. Progress will accelerate.

khafra• 14 years ago

Energy density and materials don't follow a Moore's Law curve. In 10 years, a walker this size will have similar carrying capacity and range.

Jach• 14 years ago

To be fair, they've been working on BigDog since 2005. Very fun to watch their progress but it's not exponential.

john2x• 14 years ago

Yeah. As I was watching the video, I kept thinking "why not just use a horse?".

But I think it's safe to say that this isn't going to be used just for carrying stuff. They're aiming to ultimately weaponize it and use the new tech to create other cool stuff.

InclinedPlane2• 14 years ago

Horses and camels require food, shelter, and care.

Let's say you want to have the capability of offloading 100lbs of equipment (say, 40 lbs of food, about 1500 rounds of ammunition, and about 20 lbs of miscellaneous other equipment) from soldiers so they can fight more effectively and for longer (up to a month or so), and you want to have a standing ability to be able to do this for up to 2,000 soldiers at any given time.

This means you need to buy 500 "alpha dogs" and then keep them in storage somewhere until needed.

Or it means you need to buy about 1,000 camels. Which require about 30-40 lbs of food and water a day, and produce as much urine and manure per day as well. That's five thousand tons of food and water a year, and five thousand tons of urine and manure a year. You need to build shelter to house them and you need to hire people to take care of them. You need to treat them when they are injured, you need to replace them when they are too injured to work or when they die. You need to train them. And they probably won't survive spending a month without food and water, even camels. In contrast, a robotic pack mule left without supplies for a month is merely inactive and can be retrieved and refueled.

Also, horses, mules, and camels are fragile. Far more fragile than humans under extreme conditions.

DannoHung• 14 years ago

The research version has these problems.

Go and find me another robot that can climb rough terrain hills autonomously and I'll say that the cost is unjustified.

njharman• 14 years ago

It is much harder to refine, expand, and mass produce camels. Camels manufacturing time is much longer. Camels are much vulnerable from enemy fire than metal&plastic. Camels don't have night vision. It is much harder to silence a Camel. Camels can't negotiate rough terrain. Camels need to be fed. Military logistics is tuned around delivering fuel, ammo and parts, not hay. Camels get sick, scared, injured, run away, etc. Camels spit.

Tossrock2• 14 years ago

I think there are problems with many of your points.

> Camels are much vulnerable from enemy fire than metal&plastic. Disagree. I'm quite sure if you shot one of these robots, it would stop working.

> It is much harder to silence a Camel. Disagree; under normal circumstances a camel makes about as much noise as a soldier walking, whereas this thing makes much more noise via its gas generator. This can only be silenced by going all electric, which given the poor energy density of current battery technology, is not feasible.

> Camels can't negotiate rough terrain. I am not a camel expert but I believe they can walk over terrain at least as rough as this robot.

> Camels need to be fed. Military logistics is tuned around delivering fuel, ammo and parts, not hay.

Military logistics is also tuned around delivering provisions, which are pretty similar to hay. Anyway, a cubic meter is a cubic meter.

> Camels get sick, scared, injured, run away, etc. This is the strongest advantage over a camel: precise control. You can't program a camel. However, the other points are not really big wins for the robot.

mikehuffman• 14 years ago

I have always hoped that new power and energy advances would come from these military applications.

ChuckMcM• 14 years ago

Some of the comments here remind me that not everyone watches occasional episodes of 'Weaponology' [1] :-) That show does a credible job of tracking the various technologies that contribute to the eventual weapon that you know today in the 21st century. One of the shows talked about the Tank and how it was developed to get people across the 'deadmans zone' in trench warfare and ended up becoming the backbone of mobile maneuver tactics.

So this thing demonstrates a vehicle that can navigate completely unimproved terrain carrying a useful load. In battlefields like Afghanistan there is a huge challenge getting troops resupplied (this is true in any war of course). Helicopters and other air drop means (the robotic parachutes are pretty cool) require that your troops break cover to recover their goods, these things could walk into camp where you are under cover drop off supplies.

So one tactician might air drop these things out of a plane at a safe altitude and have them land within a couple of miles of the targeted base. The vehicle starts up, and walks over to the base to drop off the supplies. It could conceivably walk back out and, using a helium balloon, send up a retrieval cable that would let a helicopter come and pull it out of there.

Thats a pretty cool capability. And in peace time if this thing could walk into a city ravaged by an earth quake and drop off supplies, that would be cool. Or carry a clean up device through a nuclear contaminated forest, etc etc. The key is 'unimproved terrain' and '400 lbs for 20 miles'.

[1] http://military.discovery.com/convergence/weaponology/episod...

DARPA's AlphaDog robot pack mule begins real world testing

Related Stories