Ask Hackaday: What Good Is A Robot Dog?

It is said that Benjamin Franklin, while watching the first manned flight of a hot air balloon by the Montgolfier brothers in Paris in 1783, responded when questioned as to the practical value of such a thing, “Of what practical use is a new-born baby?” Dr. Franklin certainly had a knack for getting to the heart of an issue.

Much the same can be said for Spot, the extremely videogenic dog-like robot that Boston Dynamics has been teasing for years. It appears that the wait for a production version of the robot is at least partially over, and that Spot (once known as Spot Mini) will soon be available for purchase by “select partners” who “have a compelling use case or a development team that [Boston Dynamics] believe can do something really interesting with the robot,” according to VP of business development Michael Perry.

The qualification of potential purchasers will certainly limit the pool of early adopters, as will the price tag, which is said to be as much as a new car – and a nice one. So it’s not likely that one will show up in a YouTube teardown video soon, so until the day that Dave Jones manages to find one in his magic Australian dumpster, we’ll have to entertain ourselves by trying to answer a simple question: Of what practical use is a robotic dog?

How Much is That Doggy on the Jobsite?

To be clear, we’re only interested in the current crop of industrial- and military-grade robotic canines. All of the robotic pets like Sony’s AIBO are interesting toys, but they aren’t intended to do the work that Spot and other robo-dogs are intended to do. Indeed, Boston Dynamics’ launch video for Spot seems to be all about the robot being industry-ready, prancing as it does through construction site vignettes with sparks a-flying and rain a-falling.

Boston Dynamics’ messaging is clear: they intend Spot to be a worksite assistant, at least at first. Its 14-kg payload capacity is respectable, but don’t expect it to be carrying around cinder blocks like the video shows. Instead, that payload will largely be made up of sensors and actuators that customers will need to haul into some sort of hazardous environment. If it’s wet, dusty, confined, hot, toxic, or any combination of the above, it looks like the company would like you to send Spot in to do the job. Applications suggested by Boston Dynamics include search and rescue for first responders, bomb disposal, confined-space entry with sensors, security, remote data collection, and even hazmat remediation.

Right now, Spot seems to be a solution in search of a problem. I think that’s actually OK, and Boston Dynamics’ approach of partnering with customers to develop the solutions to their problems is probably the right way to go. What they’ve built here is a marvelously flexible platform, and by concentrating on developing the control systems and abstracting away the complexity of the kinematics to the point where the robot can be controlled intuitively with a game console-style remote or through a high-level API, they’ve created a tool that will be able to address needs that no robotic engineer could possibly dream up.

Not the Only Dog in Town

As iconic as Boston Dynamics’ robot dogs have become, they’re hardly the only canine robots on the market. Either there’s something special about the canine layout that has lead multiple companies to pursue it for their robot platforms, or it’s a simple case of getting on the hype train started by Boston Dynamics. Either way, robot dogs seem to present a solid business opportunity, with the likes of ANYbotics, Unitree, and Ghost Robotics are all working on four legged robots. All of these companies and more are pouring big R&D bucks into robot dogs, so there has to be some mission for them. But that still begs the question: what will these things be good for?

A partial answer to that is obvious from looking at the LLAMA, or Legged Locomotion and Movement Adaptation, a dog-like robot being developed by the U.S. Army Combat Capabilities Development Command’s Army Research Laboratory as a sort of valet for soldiers in the field. Bearing more than a passing resemblance to the current crop of robo-canines, the all-electric LLAMA is designed to work alongside soldiers during field operations, carrying supplies, ammunition, or weapons that would normally be split up among squad members. Reducing the physical burden will either let soldiers operate longer in the field, or more likely, allow the entire squad to carry much more gear and provide a bigger punch with the same number of soldiers.

Your Turn

So what’s your take on Spot and its ilk? Are we likely to see robotic dogs walking around construction sites someday soon, taking measurements or bringing materials to human workers? Will fire and rescue squads someday have Spot-like robots at the ready to perform primary searches while firefighters gear up? Will there be robotic K9 police units someday, where a sensor-laden Spot will crawl under and around cars and into buildings in search of contraband or to deploy less-than-lethal weapons? Or will the military latch onto Spot and its robo-canine cousins as a force multiplier, doing the heavy work as embedded members of teams? What problems does the dog-like form factor solve better than other designs? And most importantly, for which use cases do robot dogs have the advantage over real dogs? Sound off in the comments below.

Featured images: Boston Dynamics, IEEE Spectrum (Bob O’Connor)

62 thoughts on “Ask Hackaday: What Good Is A Robot Dog?

  1. “Either there’s something special about the canine layout that has lead multiple companies to pursue it for their robot platforms, or it’s a simple case of getting on the hype train started by Boston Dynamics.”

    Four legs better than two. Extrapolate, eight might be even better.

    1. Roboticists typically want the smallest number of legs they can get away with using. Two is usually not good, too difficult to control. 4 and 6 legs tend to be pretty good options. Above 6 you don’t tend to gain advantage but do increase cost, weight, number of motors, number of motor control circuits and so-on… for each extra leg. 6 can be a bit easier than 4 as you can force a condition of always standing on three legs at any given time and hence always have a stable platform. When, like Boston Dynamics, you wish to show of your cool lifelike motions then 4 legs means you have to get them working well to operate while 6 legs isn’t quite so demanding.

    1. Not exactly the same, but I hear that Boston Dynamics collaborated on that episode. Important to note that Spot doesn’t have lidar or solar panels like the TV robot. Or, you know, explosive death spikes, a metal-cutting saw, and a malevolent AI program. At least not yet…

  2. Pets are a multi-trillion business, if they can get it to work. It is so obviously lucrative.

    Imagine no more cleaning up after it, no more germs, no more hoteling when you are away, …

    1. – Robot dogs aren’t warm and cuddly like the real thing. (yet?)
      – Total Cost of Ownership is a bit too high. US companies seem to make military/industrial types whic aren’t exactly pets. Sony has a line of robot pets that are geared towards the consumers. (pun intended)
      – Robot pets can get obsoleted without upgrade. Do you live with the obsoleted model or buy new one and recycle old one?

      On the topic of SiFi pets. A Chinese company has gone into the RePet business. They cloned cells from a dead cat and cost within reach.

      1. Boy, that book didn’t age well.

        Not that it’s bad but it seriously looks absurdly off mark about stuff. For one technology didn’t advance nearly as fast, in any field. Another thing is that US in the book is, somehow, even more insane than IRL.

        And obviously the whole terrorism thing, but we can’t blame Stross for not getting that right.

  3. “As iconic as Boston Dynamics’ robot dogs have become, they’re hardly the only canine robots on the market. Either there’s something special about the canine layout that has lead multiple companies to pursue it for their robot platforms, or it’s a simple case of getting on the hype train started by Boston Dynamics.”

    A working quadrupedal robot is probably easier to build than its bipedal counterpart. Look around, how many creatures on the face of the earth are bipedal compared to quadrupedal? Even human babies are quadrupedal for the first year of so of their life.

    1. >A working quadrupedal robot is probably easier to build

      More importantly, it’s easier to get it working. A three legged table will stand, so even if you are slow, you can still move one leg in the air and maintain static balance. Next step up is pseudo-dynamic walking, where you hold two corners and push/fall over diagonally until you catch the fall with the other two legs, then switch corners and proceed in a zig-zag fashion. Then comes full dynamic walking which calculates the positions of the legs in real time… etc. etc.

      What robots like the Boston Dynamics ones do is some cross between pseudo- and full dynamic walking. The reason the robots trot when standing still is because they’re alternating the supporting corners. When they do things like grabbing things with the arm, they go into a pre-programmed stance where the support points are known in advance and perform the motion, then return back to trotting to move around.

      It’s kind of a cheat, because they’re doing the equivalent of canned animations in a computer game, rather than real dynamic response to the environment, but it looks good enough for press and you can’t tell the difference if you don’t know what’s going.

  4. For ways that robot “biomimetics” can be used, at least by a paramilitary organization, I highly recommend the book “The Last Good Man” by Linda Nagata. It’s a thoughtful near-future SF with lots of action and very cool projections of how robots will aid and supplant soldiers, both for the tasks a human would do, and to keep wars running for the multinationals who profit from them. https://mythicisland.com/the_last_good_man.php

  5. SF author Phillip K. Dick pretty well addressed that in his ‘Will Anderoids Dream of Robot Sheep?’. His novel was the inspiration for ‘Blade Runner’ and ‘The Man in the High Castle’.

  6. A construction helper? Worthless. This is what apprentices and day laborers are for. They are much cheaper than a $100k robotic toy designed by autistics. I can’t see any construction company except the most wealthiest buy this junk that costs as much as new bulldozer or couple of BobCats.

    A pack mule would serve our soldiers better than this POS. Any robot that can carry a large load will require a equally large number of technicians to maintain. And if it breaks down on the battlefield you are SOL.

    Also the “dog” looks like a nightmare only a autistic could love. I can see terrorists buying them and turning them into suicide bombers to blow up malls filled with people. Or some hacker accesses it and makes it attack it’s owner.

    A drug or bomb sniffing dog substitute? I’ll take a real dog.

    1. What the actual fuck is with the vitriol towards folks on the Autism spectrum here?

      Pull your head out of your ass. It’s a robot, it’s cool, and it’s miles better than options we have now for semi-autonomous mobility.

    2. What exactly do you mean with: “designed by autistics”
      What exactly do you mean with: “looks like a nightmare only a autistic could love”
      Then you mention something that is obviously not very nice towards terrorist and hackers. The hacker thing I can imagine, but the terrorist buying one seems a little over the top, considering that this things costs a lot. Perhaps in a few decades or so, then the prices have dropped and the second hand market has kicket in. However, it will by then be almost impossible to buy a replacement battery that actually fits. Because for some reason battery packs are always a different size. Somehow they manage to design a completely interchangeable batterypack for every kind of laptop created over the past 3 decades. While the shape of the laptop almost never changes. How difficult can it be to design a standard battery casing. The cells inside are always the same size. Why not a battery compartment, where you slip in 8 or ten cells, just like the good old ghettoblaster days?

      The POS wasn’t very clear to me, did you mean: Point Of Sale, Plan Of Study, Port Of Seattle, Power Of Sugestion ?
      The SOL also wasn’t very clear, did you mean: Standards Of Learning, Sons Of Liberty, Serial On Line, Software Object Library

      Regarding this robot dog, I think it’s marvelous, not for the concept of having a dog or replacing a dog, but for a small platform capable of carrying sensors, sensors capable of searching for landmines, sensors capable of search and rescue missions (perhaps a Dalmatian themed model for the fire brigade). For entertainment purposes the could be used to held races. Because people like to race anything that moves. Perhaps racing robots could be the next best thing to fighting robots. Or perhaps a combination.

      Anyway, what I mean is, the future can be bright, just be prepared to believe in it.
      Because, for anything you do, there is always a reason not to do it. Think positive.

      1. Autistic people have the tendency to ignore other opinions because they have a reduced capacity for forming a theory of other minds. That is, the first instinct is to assume that everyone else has the same information, opinions, and feelings about the surrounding reality as the autistic person. That tendency also applies to their own past self, so there’s a lack of self-criticism because reflecting on past mistakes requires the imagination of themselves as a different person than they currently are.

        In design, this results in the stereotypical “engineer’s interface” which is obscure or completely inaccessible to other people, which the autistic person finds intuitive because they have designed it, and they can’t imagine why other people would find it difficult or cumbersome.

        “Autistic” as a word derives originally from “auto” and “-ism”, and its purpose was to describe a person who is not necessarily egoistic nor narcissistic, but self-absorbed or “self-sufficient” in the sense that they find no need to engage with other people on a social level. A shut-in, intellectually or otherwise.

        1. First, to quote the overused phrase: “Autism is a spectrum”.

          Second, “Autistic people have the tendency to ignore other opinions because they have a reduced capacity for forming a theory of other minds. That is, the first instinct is to assume that everyone else has the same information, opinions, and feelings about the surrounding reality as the autistic person.”

          By that you mean the Dunning-Kruger effect? You know, get so good in a field you start thinking everyone have the basic knowledge on that?. Because if so, pretty sure it apply to way more people than just autism.

      1. My son actually had a good use case for this: make Spot (or more likely LLAMA) lug a heavy machine gun and ammo. It could be maneuvered into position and act as the mount for the gun, allowing the human gunner in the squad to actually run the gun.

    1. If you want to strap weapons to it then heavy support weapons make the most sense. Think mortars, anti-tank missiles like the TOW, things that are close to or just outside of human weight capacity. Give the soldiers an aiming device to slap on their rifle. Instead of being burdened with a heavy Carl Gustav you get a sub-1lb laser pointer of death and the robo dog does the work

      1. also the robo mule can carry sensors that are too heavy/bulk for a helmet mount. Thermals with magnified optics are frigging amazing. The interface for that is crappy now but eventually AR/VR will be usable in the field. Then soldiers can get thermal overlays from the robo mule.

  7. I would have loved a pair with thermal heads for search and rescue in burning/smokey buildings. We were first getting shown IR helmet visor things for everyday use(forest service had had them for a while earlier with cryogenic gas cooling) when my body got wrecked at work.
    Give me one or two strong enough to drag a 100kg firefighter+gear or drag a patient out and you have one of the best use cases I can imagine outside the most intense warlike situations.
    I want robots to get broken for the public good, no more human firefighters and other first responders who then get denied benefits when they cant work anymore. You can bin a broken robot, but both breaking and then throwing away human firefighters is something we need to get past as a soceity.

    1. My thoughts exactly when I retweeted the article link? Can it drag a 200lb person from danger? Can it carry the gear for a special needs child? They can open doors, so have them open doors for those that can’t. Can it be trained for safety? Street edges, hazards, etc Can it replace service animals? The 90 minute run time actually limits a lot of military scenarios, even with replaceable batteries. But can it patrol my property at night and recharge in between rounds?

  8. If it can go to the refrigerator, open it, grab me a beer then bring it back to me unscathed, then it’s already more useful than an actual dog.

    Plus, it won’t annoy my neighbors with relentless barking, stink up my house, crap on my floor, ruin my belongings or beg for food!

  9. There are reasons that nature created creatures with 2, 4, 6, and 8 legs… and then a few segmented creatures with many more than that, and some with 0, and even some with just a spiral tail. Would hesitate to discount Mother Nature, Best to just study what she’s laid before you now that we have 3D printers.

  10. Quadruped motion could be hugely useful. Once it is engineered to the point of being reliable (and fast) it could make personal transportation four-legged. The net result would be no need to cover the ground with vast quantities of concrete. A continuous constellation of stepping stones would do nicely and if the foot-pads are big enough, even that would be unnecessary. We could literally un-invent the wheel and let our four legged transport with head shaped add-on chomp down and “swallow” the user and place him in his pilot’s seat inside. After traveling to one’s destination it could just open it’s “mouth” and let you out. …or if the weather is good you could just ride on top like David the Gnome rides Swift the Fox. …now that would be trippy, better than riding any motorcycle or in any convertible. Yea, robot dogs are very useful because they open the door to lots of quadruped uses. Until then, just having them walking about the house doing minor errands would be great. ….”Hey dog-bot. There’s no toilet paper in this bathroom. Be a good dog and bring me a roll from the closet, please.”

  11. The Benjamin Franklin quote reminds me of a quote from Michael Faraday, who invented the electric motor, the electric transformer, and many other useful things. One day the Prime Minister of the time visited him at his laboratory in the basement of the Royal Institution, and having seen his work asked what good it was. Faraday replied “One day you will be able to tax it!”

  12. The main limiting factor, is always the power source. The more work you place on these platforms, the great the fuel consumption, just like anything else. I’m guessing the battery last just about long enough for a short video, to get people interested. They will always be cool toys, until we find a better energy source to keep them powered for hours, doing work.

  13. I work in a Micron clean room, and Id use the shit out of one of these.

    I could send it to fetch parts and tools,

    I could have it go preform inspections on equipment

    hell I could go have it run a batch of wafers in a carrier to the other side of the factory without messing with the overhead transport system.

    i want one badly!

  14. I think that there would be tremendous applications for these on the moon or Mars. The cost of keeping humans alive and well on other worlds is very risky and expensive. With a limited number of humans to make all of the decisions of even simple tasks like pulling wires or pipe from one location, or module to another. Humans in space suits doing hard labor is just not a good idea. Robots don’t need air or water, just recharging and maintenance.

    Imagine setting up the first modules on the moon to allow for colonization and extraction of water ice to produce air, water and rocket fuel. Will humans in pressurized space suits be using pick axes and shovels? Not a chance! The modules need to be set up, wires and pipes need to be run between them, and ditches will need to be trenched to protect those wires and pipes from rovers wheels and cosmic radiation. Solar and radiator panels need to be placed and set up. Storage for water, air and rocket fuel need to be set up. What could be better to do some of these tasks than robot dogs with cameras and hands?

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.