Sunday, June 14, 2015

Big Dogs and Little Horses

An infantryman in the US Army carries round 100 pounds of gear into the field with him. The Army would like to off-load that gear onto robot. For a number of years now, Boston Dynamics has been addressing the problem with a number of generations of a robot including Spot, Little Dog, and the famous Big Dog, as well as its successor, Alpha Dog.

What Boston Dynamics have managed to do mechanically is astonishing. Here's a video introducing the lastest, "Alpha Dog", known to the military by the obligatory yawn-inducing designation of "LS3" (Leg Squad Support System):

Boston Dynamics (BD) and the DARPA both describe Alpha Dog as a "robotic mule". This got me wondering, how do BD's robots stack up against real mules? After all, if you're trying to replace a thing, you might ask whether your replacement is better than the thing itself, right? I looked at some specs published by BD (PDF) (I had to extrapolate some) and did some quick math. It's not going to be exact, but it should do the job. My info about the mules came largely from having had a stepfather who owned a farm, plus a search or two to confirm some figures. This ain't an exhaustive comparison, folks. Rather, it's me thinking out loud.


A new BD robot can cost as much as $32 million to develop. A production model might cost $100,000 dollars. A mule typically costs between around 2 thousand dollars to buy, and about the same per year for food and maintenance. In comparing operational costs, I exclude the cost of military labor: these guys get paid regardless.

For perspective, you can own, feed, and shoe a mule and provide him with regular veterinary visits and buy replacement mules for over 40 years for about what you'd spend in acquiring one Alpha Dog. This disregards the development costs of the robot. It also doesn't take into account what you'd spend on gasoline, replacement parts, and highly technical mechanics (the analog of veterinarians). This is mainly because you'd only need to gas up a robot when you're actually using it whereas a mule eats every day.

An Alpha Dog carries twice what a mule does, though, so you'd only need half the number.

We're going to have to call this in favor of the mule. You might argue the mule is more expensive than stated because of the need for a stable and paddock, but I'd just throw the  $32 million development costs of Alpha Dog back into the equation and more than even things out. Best to not push it. Also, my graphs show 5-year costs because in robotics you're constantly working on making the current model obsolete. This isn't a jeep: they'll go obsolete frequently. And robots aren't replaced kind-for-kind. Remember the cost of a new mule is a few thousand. The cost of a new model robot is more R&D... to the tune of millions of dollars.

Mule wins, by a long shot.

Relative costs minus R&D. Concerning operating
costs, I think I'm being kind to the robot.

Relative costs, including R&D. I included the cost of buying
a paddock and stable for the mule as "Development Costs".

However, the tools of war aren't chosen on cost alone, or we'd all still be fighting with slings and arrows and flying Sopwith Camels. Let's keep looking.


A mule can carry around 200 pounds of cargo. The Alpha Dog is rated at 400 pounds. So one robot should do the job of about two mules. Also, an LS3 model would appear to go pretty much anywhere a mule could go. Go Dog!


A mule gets tired. An Alpha Dog doesn't. This seems an advantage, until you remember that this robot is designed to walk along with men on foot, and they get tired, too; faster than mules.
"A difference which makes no difference is no difference at all."

Everybody's on foot
Public Domain image via


Keep in mind the words of the briefing in the preceding video:
"...squads want to move under cover and concealment. They do not want to move on roads. That tells you right away that if you can have a robotic platform, it must be able to go over logs, through streams..."
Yeah, but in order to meet that requirement for "concealment" it would have to go over logs and streams and rocks, and all the rest fairly quietly. Now have a listen to a video that hasn't had the sound muted for voiceover or catchy music:

One of the things you don't get to experience in many videos of the Boston Dynamics "dogs" is how incredibly LOUD they are. This is because they run on electricity supplied by a generator powered by a two-stroke internal combustion engine as well as hydraulic actuators powered by the same. This thing is like firing up a chainsaw, then revving it constantly as you walk through the woods.  In the lab they're a lot quieter because they're tethered, with hydraulic pressure supplied through an umbilical.

A mule, on the other hand, doesn't have that much to say. It might bray now and then, but it never gives off a constant 110 decibel advertisement of its position. And while a mule does give off an odor, so does the gasoline exhaust of the robot.

The mule wins again. .

Portability And Storage

An LS3 model can be packed in a crate for years and airdropped into a combat zone at need. Mules need constant care and feeding, they take up more room in transit, aren't easily stacked, and tend to die horrifically if you throw them out of aircraft.

No contest. The robot wins.

Other Factors

If your Alpha Dog is out of gas, it isn't moving until refueled. That's no different from any other piece of mechanized equipment. However, if your mule is hungry, it can graze. Of course, this is no solution in an arid desert, but even most deserts aren't that arid, but in that situation you have no more chance of finding gas than grass. Foraging is foraging, whether it's searching for a patch of green or siphoning gas tanks. Either may have the advantage depending on where it's deployed. No winner.

An Alpha Dog may be more resistant to small-arms fire than a mule. If you're facing an RPG attack, I don't see that either would fare well. However, a mule has a sense of self-preservation whereas a robot does not. You might be faced with gunfire to find that your mule has turned tail and run away. Advantage robot.

I have to put some consideration as to what I expect the pack "animal" to carry. It's been a long time since I was in the military, but I'd want to carry my own small arms and as much ammunition as I could whether I had a mule, Alpha Dog, or a Radio Flyer wagon. The things I'd off-load would be stuff like tents, bulk food, additional ammo, and bulky and heavy specialized weapons, etc. And while I'd want something that will stick close should the battle come to me, I think I'd be unlikely to deliberately take an LS3 into any situation where the din of its gas-fueled power plant will give me a disadvantage. Besides pinpointing a squad's position, the sound would prevent them from hearing the enemy.  Best to leave it behind with a guard or handler, just as you would with a mule, and turn it off.


Boston Dynamics has been performing wonders with their military robots. Based on its maneuverability, survivability, and carrying capacity, as well as its ability to be easily stored and deployed, I can see why the military wants these things despite their obvious expense. Expense, after all, has never been a huge problem for DARPA. I'm sure they think otherwise, but all things are relative.

Nevertheless, I can think of some significant downsides to the robots in some situations; and I'm certain that the various armed forces are well aware of all of them, and more. Boston Dynamics are constantly addressing these concerns with additional features like manipulator arms and greater autonomy. I can also think of situations where the mule is arguably better, but that argument is long past. Cavalry is mechanized. But who knows... maybe someday we'll have a robot advanced enough to recognize danger and run away.

I also think that the applicability of this robot is somewhat overstated even while its capabilities are essential in the circumstances for which its designed. The reason for using a legged robot is to maintain cover in difficult terrain. In that very situation, the din of this device is a major drawback. In most situations I'd say that not only would a Humvee or wheeled "mule" do the job, but it would do it better with more carrying capacity. So I doubt you'd see one of these robots in most situations. In practice you might carry an Alpha Dog as cargo in a truck should your mission require it, deploying it on legs only when needed.

Finally, war is the most evolutionary process we have. By this I don't just mean that it spurs technical innovation, though it does. Rather, I mean that it's survival of the fittest, not the most advanced. Wars are won more by force than showmanship; and highly trained, expensively outfitted soldiers have been taken out with nothing more advanced than a pipe bomb in a baby carriage. While it's impressive to have a high-tech robotic "Gunga Din" along, it must continue to be human brains and mission planning to determine whether it's worth its bulk.

An M274A4 4WD Mechanical Mule.
These have been replaced by the M-Gator from John Deere
Photo via

No comments:

Post a Comment