Junebug-like robotic pack mule

Here's a DARPA video showing a robotic pack-mule prototype. I think you're supposed to imagine this thing being on your side, but when I see videos like this, I always find myself imagining what it would be like to be crouching in the underbrush with a couple of terrified children, trying to keep them silent while this thing motors through the uncanny valley around us.

This video depicts field testing of the DARPA Legged Squad Support System (LS3). The goal of the LS3 program is to demonstrate that a legged robot can unburden dismounted squad members by carrying their gear, autonomously following them through rugged terrain, and interpreting verbal and visual commands.

DARPA Legged Squad Support System (LS3) Demonstrates New Capabilities (via JWZ)


    1.  Yeah, I was going to say this reminds me of the Romney’s dancing horse.

      The high stepping is an inefficiency because real animals (and humans) usually don’t lift their feet any more than necessary to clear the ground.

  1. Scaled up the noise of a junebug.
    I’d have thought they’d have found a better approach than this feeling its way
    prancing / dressage business.
    Video obstacle identification; kangaroo (rat) – like jumps; it’s so falteringly timid as developed, like their ‘dog.’

    1. Let’s try powering that with batteries or a fuel cell if that will provide enough horsepower/endurance and cut out the noise. You’re not going to sneak up on anyone except deaf people with that thing.

  2. I think it’s prance-y walk is cute, actually. Add bunny ears, and I’d adopt one myself! Unfortunately, it’s darned loud. I can’t imagine wanting to go on a hike with a mobile generator buzzing away in my ear.

  3. Maybe they could even be used to provide improved mobility for paraplegics, if they don’t work out for… you know… helping us kill each other.  NAAAAH – what was I *thinking*?

    1. Oblig Simpsons:
      [At the Geriatric Olympics]
      Grandpa: Molemann’s leg prostheses give him an unfair advantage!
      [moleman in a pair of 10 foot robot legs jumps over the high jump bar set 20 feet high]

  4. Just seems so fussy and full of things that will need maintenance and babying along.  What’s wrong with using mules…they were bred for it and have gone to war for generations…I wonder if it is because there are so few farm boys/girls anymore; no one knows how to handle stock?

    1. You can’t use a mule to crush a bunch of people to death.  You can’t direct a mule to autonomously run into battle and offload a 20mm CIWS turret and power it with the generator.  You can’t load a half ton of freight onto the back of a mule.

      And most importantly, you can’t turn a mule off to save gas/food.

      You also can’t pack mules into a C130 allotting only half an inch of space on each side between animals.

      Basically you can’t treat mules like shit, since they’ll die.  Once this thing is ruggedized and mass produced, it’ll probably be as durable as any other army vehicle.

      1. Well, huh.  Good points all.  I was going to say, sounds like we are preparing for combat in rugged terrain, for the long haul.  But then realized the features you mention would be valuable in urban warfare as well.


        1. In urban warfare you can use wheels or tank treads, which are many times simpler and more efficient.

          I think this is definitely for rugged terrains.

      2. I find your logic default,
         1.You could probably train a mule to carry machinegunturrets with a lot of patience and a little love
        2.Mules dont need to run on gas, they run on grass and hopps, cheaper than oil and more abundant.
        3.You could very well pack mules inside a C130 with only a 1/4 of an inch between them accepting that a lot would not survive (or even tighter if all were dead).
        4.Mules are treated like shit all over the world, all of them die but only some sooner than others.

        I find mules quite rugged creatures and smart in having hidden theyre war-machine like talents for so long. But no longer.

        Also, rhinos would kick Junebugs ass!

  5. Chapter III The Death of All Men

    Carlotta stared at the machine. It had legs like a grasshopper, a body like a ten-foot turtle, and three heads which moved restlessly in the moonlight.

    From the forward edge of the top shell a hidden arm leapt forth, seeming to strike at her, deadlier than a cobra, quicker than a jaguar, more silent than a bat flitting across the face of the moon.

    “Don’t!” Carlotta screamed in German. The arm stopped suddenly in the moonlight.

    The stop was so sudden that the metal twanged like the string of a bow.

    The heads of the machine all turned toward her.

    Something like surprise seemed to overtake the machine. The whistling dropped down to a soothing purr. The electronic chatter burst up to a crescendo and then stopped. The machine dropped to its knees.

    Carlotta crawled over to it.

    Said she in German, “What are you?”

    “I am the death of all men who oppose the Sixth German Reich,” said the machine in fluted singsong German. “If the Reichsangehõriger wishes to identify me, my model and number are written on my carapace.”

    The machine knelt at a height so low that Carlotta could seize one of the heads and look in the moonlight at the edge of the top shell. The head and neck, though made of metal, felt much more weak and brittle than she expected. There was about the machine an air of immense age.

    “I can’t see,” wailed Carlotta. “I need a light.”

    There was the ache and grind of long-unused machinery. Another mechanical arm appeared, dropping flakes of near-crystallized dirt as it moved. The tip of the arm exuded light, blue, penetrating, and strange.

    Brook, forest, small valley, machine, even herself, were all lit up by the soft penetrating blue light which did not hurt her eyes. The light even gave her a sense of well-being. With the light she could read. Traced on the carapace just above the three heads was this inscription:


    And then below it in much larger Latin letters:



  6. It’s fascinating to see what works.

    For decades people have wondered about the best forms of robots for different tasks; humanoids for serving beer, one armed behemoths for spot welding and casting. Now we get to see what works on Mars, what’s the fastest pick-and-place food processor, and what works as a pack animal on earth. Some hydraulics, a generator and assorted do-dads and this is what works. Rubber feet, simple lidar, some linkages and kinematic algorithms and here we are with something that works. This is about us, walking backward over rough terrain with a video camera.

    I’m optimistic that even simpler robots could defeat the next few generations of this type of war machine. Better applications will come along that we can feel warm and fuzzy about. Search and rescue in whiteout conditions? Then I wouldn’t mind hearing the whine of a few machines.

  7. I love the robots these guys make. They’re very impressive. Thing is, and I hate to be a, I dunno, party pooper, I don’t think I’ve ever seen one of their robots doing anything on four legs that wouldn’t be better accomplished on wheels or treads. 

    But then, I guess the scientific advancement is more important than the application. I mean, I remember my jaw dropped so hard on seeing that one video of  that guy trying to kick Big Dog over on an icy parking lot only to have it keep it’s balance and get right back up that I had to climb under my desk to find it. I’m sure that can have applications outside of an impractical pack bot.

    1.  moving over rough terrain, going up steep inclines(on lose soil as seen in the video at that incline the wheels or treads would just spin while the bot rolled down the hill), maneuvering around trees and brush (maneuvering on treads is terrible) also moving on ice (wheels even treads will sit there and spin, big dog took it like a champ)

  8. I dunno. Back when Big Dog made the rounds I used to worry about the development of robotic weapons. But more and more I see them as terribly inefficient and unfeasible.

    Mechanized weapons platforms are frightening, and for good reason. But ultimately machines are not (and for some time will not be) flexible or adaptable enough to deal with focused human resistance. It’s one thing to send a UAV on a long range bombing mission with little to no chance of interception. It’s another entirely to send it into a dogfight, or through a corridor of SAM batteries.

    A mechanized pack mule may offer a certain measure of benefit when you’ve got boots on the ground in rough terrain. But just how much benefit does it really provide, and in comparison to how much detriment or cost?

    The machines we’ve seen so far are pretty crude. Big, noisy, brainless, and so far unarmored. We can of course assume some level of improvement in all those areas down the road, but for now let’s assume they’ll be at least a mild nuisance to have with a unit of troops. They’ll make it more likely to get you spotted; they’re at least somewhat prone to sensor errors, software glitches, or mechanical failures; and they might as well have a big bullseye painted on the side with chunky block letters saying “Shoot Me, Please!” because their heavy burdens of equipment will make them attractive targets to disable, consequently bring their platoons to a halt with a large supply of now immobile equipment.

    How much will it cost to create enough of these things that the ground units that actually would benefit from them will be able to make use of them? How much will it cost to armor them such that they can’t be immobilized / blinded / destroyed with a few well placed bullets or an explosive? How much will it cost to fuel them, to repair them, to maintain them? How much will it cost to environmentally seal them to prevent failure due to humidity / extreme heat / extreme cold / sand and soil / magnetism / et cetera? How much will it cost to secure them against electronic warfare such as signal jamming / command spoofing / data interception / et cetera? How much will it cost to recover damaged or immobilized units, or to replace destroyed ones?

    Ultimately the big question that comes to my mind is this – wouldn’t it be cheaper and just about as effective to simply use ACTUAL pack mules? Or would the thought of animals suffering be too much of a blow to the public relations side of the equation?

    1. Hmm…let’s see:

      “The ‘flying machines’ we’ve seen so far are pretty crude. Big, noisy, brainless, and so far unarmored. We can of course assume some level of improvement in all those areas down the road, but for now let’s assume they’ll be at least a mild nuisance to have with a unit of troops….they’re at least somewhat prone to mechanical failures; and they might as well have a big bullseye painted on the side with chunky block letters saying “Shoot Me, Please!””

      We’re looking at a prototype here. I’m quite sure these will become a ubiquitous part of our world, both military and non, in a not-too-distant future. Like ‘flying machines’ did…

      1. You realize I’ve been hearing “these will become a ubiquitous part of our world” for 30 years now.

        Our “flying machines” however cover great distances and do things that people/animals can’t do on their own.  This, not so much.

        And in terms of the evolution of aviation, the first 30 years make some pretty big strides.

        1. From who have you been hearing that, other than science fiction?  My father, who worked with the military in weapons development, told me thirty years ago that airplanes would eventually become obsolete as a primary weapons platform, when we would use drones with no pilots that could fly faster than humans could tolerate and would kill themselves without hesitation.  He didn’t use the term ‘drones’ because that wasn’t in circulation in 1982….but that’s what he was describing.

          Robots HAVE become a ubiquitous part of our world….most people just don’t recognize them because they’re not anthropomorphicized or are so dedicated to one function we think of them as simple machines.  When I was a kid, robots building cars was the stuff of science fiction, not fact.  Military units going into battle with UAVs is not standard stuff, drones serve the military in a host of applications.  Sending in robots to remove mines, defuse bombs or enter a fire-zone to dismantle defenses is becoming common.

          There was a time when tanks were considered unwieldy, too expensive and too clumsy to be used in practical warfare.  Could this thing be super-vulnerable?   It certainly looks like it. So were tanks, once.  A mobile all-terrain self-impelled platform that could carry additional weapons, ammunition, electronics packages, medical supplies and also act as muscle, a towing device and who knows how many other applications?   I can easily see the lessons learned here paying off.

          The problem with saying that the first 30 years of flight had big strides is getting everyone to agree when that actually WAS.  If you mean from the first powered flight by the Wright Brothers, then yes.  But by the time THEY did it, people had been attempting to build similar devices for 100-150 years.  

          Right now I’m reading Erik Larson’s “Thunderstruck”, which is partly about Marconi’s invention of radio telegraphy…and it’s stunning how in 1901 no one can see how it would be useful or how in 1905, when transatlantic transmissions are being tested, no one can see it having broad applications or have any idea how dramatically it will be adopted, less than 30 years later.

          1. The thing with radio is, people saw applications for it as a telegraph to send messages where there were no wires: ships at sea, that sort of thing.  What people really didn’t foresee was broadcast radio, the way it enables one to many transmission.
            nb the old term for drones was RPVs (remotely piloted vehicles)

          2. They did forsee one-to-many transmission…they just didn’t understand the application usages, at first.  Wireless was slower and wires had already been run.  One of the PROBLEMS of wireless was that it was one-to-many.  In fact, one competitor totally pranked Marconi’s company at a presentation by proving that it wasn’t as secure with this new ‘tuning’ technology as Marconi had implied.  Everyone was afraid that messages wouldn’t be private.  Just like no one grasped to power of Twitter now, they didn’t grasp the application of radio telegraphy at the outset.

            The point being, IMHO, that technological adaption can be rapid when the utility emerges.  In 1993, people considered the web a curiosity, but something only fringe nerds would need or could use, but a few short years later it was ubiquitous. (I remember how many people got news about 9/11 over e-mail, the web and forum postings, fr’ex.)
            What hamstrung it’s adoption and development was the requirements of both the governments involved and the commercial environment in which it was fostered (as well as the scientific community’s politics and dynamics at that time).

          3. id like to point out drone strikes pretty common, they have a form of drone strike called a signature strike, where they kill someone with a drone based on some loose pattern matching, like he was going to a place frequented by terrorists or carrying something loosely shaped like an RPG. a drone can carry a bomb and take out an entire girls school(well the fire did most the work)
            whether its spying, assassination, attacking known terrorists or just attacking suspected terrorists you know nothing about, drones serve a strong presence in combat.

  9. The most interesting thing, to me, about these robots is how they are programmed. Instead of the old AI approaches, where control and decision-making was top down & centralized (and getting a robot arm to pick up a particular cube was a real accomplishment), here, the “intelligence” is in the individual components, which gather data about the local situation and respond appropriately, with no central program “telling” them what to do. Rodney Brooks was an originator of this approach (he also happened to be my LISP instructor, but that was much earlier…)

  10. It would not be hard at all to add a standard weapons turret with an M240 to this guy and have a remotely operated/fired weapon, or the weapon could be switched over to  search/track with the fire/no fire option being held remotely by a person… or just turn it loose and it will fire at anything that moves (Israel has had automated turrets for years guarding their walls)

  11. This is really cool. That first Boston Dynamics pack mule that showed up on web a few years ago blew my mind.

    Although this is a long way from actual use on the (battle)field, If ever.
    And all they have to do is leave one malfunctioned robot behind, and it will be a waste of years of research and $$$$$$ once the enemy reverse engineers it.I still wonder how stupid people feel about spending a jillion US$ on R&D for that stealth drone only to have Iran take it over and land it intact. Of course with stealth technology, its pretty much obsolete once its deployed. Maybe it will be the same for these DARPA pack mules… By the time they actually start using it, there will already be new and better ones in development.

  12. “Boston Dynamics”.

    I suppose this explains the name of the Boston company “Massive Dynamics” that’s featured in Fringe.

  13. In Firefly, the use of horses as beasts of burden on backwater worlds made perfect sense, even with the fact that they have to eat and sleep and whatnot. Horses can make more horses, scavenge food from a dynamic, unstructured environment, and self-repair with minimal intervention from humans. I have yet to see a dune buggy or robot do any of those tricks.

    However, that only matters if you are in the middle of nowhere with effectively no supply lines, and are expected to stay there for the rest of your life. I think the US Military doctrine in that case is to call for backup and get the hell out.

    Assuming you have supply lines, the fact that these things run on gas and can’t self-repair isn’t all that important, as even if they are “attractive targets to disable, consequently bring their platoons to a halt with a large supply of now immobile equipment”, the response is to have someone strafe the area that the disabling shot came from, and then have the supply line bring up repair and replacement matériel, not sit around with your thumb up your ass going “Well boys, I guess they got us, time to call off the war”.

  14. An autonomous vehicle equipped with all-terrain treads will get your cargo there faster, cheaper and a hell of a lot quieter—but it just won’t haunt your enemies’ nightmares in quite the same way.

  15. I’m guessing the main reason Darpa aren’t relying on actual pack mules is that they figured their Terminators would bond better with robotic pack-mules. Companionship gets a lot of people through war.

Comments are closed.