As Design News' Rob Spiegel reported last month, Google is developing a robotics division. After buying up robotics companies during the last several months, the company made this announcement in a New York Times article. But 10 days later, the Times revealed that the company had purchased Boston Dynamics.
Those are the geniuses who pioneered animal-like robots based on emulation of animals' actual movements, such as Big Dog and Cheetah, robots that move like bugs, and the humanoid stair-climbing Petman. The company has been under contract to DARPA since it began: most of their robots were developed for military uses. Currently, it's developing humanoid robots for the DARPA Robotics Challenge.
Google's expected target applications for its new robot division are in manufacturing and retailing, and its other robot purchases are right in line. So why did it buy Boston Dynamics, makers of innovative Big Dog, shown here, and the leading-edge military robot company?
(Source: Boston Dynamics)
I'm not surprised that Google wants to get into robotics. It's certainly an area of technological growth -- in fact, multiple technologies. But the first Times article reported that Google's expected target applications are in manufacturing and retailing. Its previous robot purchases are companies that specialize in technology you'd expect to find in that environment, like robotic arms, grasping hands, and machine vision.
So why did Google buy military robot technology? Boston Dynamics' robots aren't little four-wheeled baby tanks, either, like some military or search and rescue bots. They're big, strong, rugged, autonomous machines built with sophisticated and sometimes revolutionary technology. The company's founder, Marc Raibert, is a former professor at Carnegie Mellon University and MIT, and considered by many to be a robotics visionary.
The fact that the Massive Dynamic of this world (a Fringe reference) -- a company said to have more computers than any other -- wants to use military robot technology sets me back. I'm not the only one that finds this a puzzling, even unsettling, move. As is usual in much of the press on robots, some coverage of the Boston Dynamics purchase was accompanied by jokes about our robot overlords and mentions of the Terminator movies. The funniest, I think, was a blogsite called Bustle. The headline for that blog reads, "Google Buys Boston Dynamics, Will Soon Be God."
The blog that echoes my thoughts the most was Forbes' Robert Hof. He cited Google's extensive information gathering on its users, its huge computer network, and its new artificial intelligence team, as cause for concern when combined with autonomous military robots. That part appeared to be a joke. But like him, I wonder what Google intends to do with this combination. It seems far too high-powered for the factory, and a curious business move. I'm not at all sure this is a good thing.
Google may have plans to offer a replacement for letter carriers, robots that won't go postal. That might save, in one fell swoop, the USPS and Saturday mail delivery and also eliminate further growth of their mandated retirement fund. Robots don't require pensions when they retire, or do they?
I would guess the likely reason is there is something in Boston Dynamics' code that google is interested in. Balancing on legs while moving takes a lot of processing and Boston may have a more efficient means of doing that with minimum computing power.
It could also be that they have so much money they can afford to own something just for the "cool factor" of it.
I'll worry more when the Google company motto is no longer "don't be evil".
Good point about code, JimRW. That's certainly part of how BD managed to get more animal-like movements in its bots. Google has shown it doesn't operate by its own motto in several different ways. That's one of the reasons it reminds me of Massive Dynamic. Some cynics would say that's why it adopted that motto in the first place.
Here's what I think, based on Google's (and others) activities:
A huge market is in-vehicle info-tainment systems. The Automotives, by and large, all have crappy systems, although some people except On-Star. The future of in-vehicle info-tainment is going to be plugging in your phone (wirelessly, of course) and making use of its applications, bandwidth, and constant updates. The Automotives will just supply the cradles: large screens, easy connectivity, standardized safety interlocking for distracted driving.
The problem for Google is that they are late to the game and well behind others, like Microsoft and Apple. Google doesn't want to be just a me-too company, so they are looking for an advantage. And that is why they started their autonomous vehicle program: They want to define the future for vehicles.
A fully autonomous vehicle is a difficult prospect. It's easy to use robots when decision making is limited; but the driver of a car needs to be able to evaluate multiple information streams and make decisions quickly, especially in an emergency situation. So certainly, Google would be interested in autonomous robots. They would be especially interested in autonomous robots like Big Dog, which can climb uneven and varying terrain, AND react to unexpected events, like a tremendous shove(very impressive). I would also expect Google to be interested in partnering with DARPA and companies like Lockheed, who are working to develop pilot-less fighter jets.
In other words, I think it's long term move that makes commercial sense. There's also the potential side benefit of unexpected synergies that can take off in a totally new direction.
Google, that company that gathers up so much information about everybody who uses the internet, almost, is now investing in some serious robotics, and has purchased an organization that has a proven track record of producing "things that work well", and nobody is concerned? Possibly GPS guided versions of "Big Dog" could be used for delivery purposes without many of the limitations faced by the Amazon Drone Delivery System. Just consider that a thing walking down the sidewalk does not need to meet the requirements for an autonomous motor vehicle, nor does it fall under any FAA rules. So a Delivery Dog system would be a lot less regulated, and a whole lot "cooler" than the delivery drones. But then I think about those "hounds" in the book "Farenheit 451", which seemed to be able to complete a task without intervention. Not all uses of autonomous devices would be for the benefit of most of us. I can certainly imagine being followed by a robot of some kind delivering an advertising spiel for some shop that I was approaching on the avenue. That would be one step beyong the location sensitive advertising planned for smart phones, and much harder to ignore. And much less pleasent yet is the image of a robotic police droid enforcing the latest whim of an overbearing city government. That may sound way outside the realm of possibility, but it is almost within the scope of current technology. And who believes that all advancs in technology are "for the good?"
Too much ground clutter, traffic, to use robotic delivery dogs in most urban locations. And the power plant can be quite noisy too and less energy efficient than, for example, a lighter than air blimp drone. I can imagine, even if the robot's power system was relatively quiet, its animal like shape and movement might excite a lot of barking dogs.
The code required for dynamic balance control of a quadruped isn't particularly applicable to the act of driving a car. However, obstacle avoidance, and maintaining a correct route are activities whose clever coding could be valuable to Google's autonomous vehicle persuits.
Buying into the expertise of Boston Dynamics could be very helpful to Google. And yes, it does give one pause to think about the modern day goldrush to centralize capabilities into huge conglomerates. On the other end of the cycle, a time may come when pressure will come to bare to once again disassemble the monopolies. Human history seems to repeat itself with no promise of learning from it.
I think bdcst is right about using Big Dogs for mail delivery. That's kind of overkill: they're huge and expensive and made for difficult terrain, not exactly a description of urban or suburban landscapes.
Thanks for your comment, William. I think some people are concerned, as I noticed in much of the tech press reportage on this. But that's only journalists and bloggers. Whether lots of people are concerned is another question. And even if they were, what could they do about it?
Hi Ann--It is a curiosity why Google would pick up the diversity of companies it has acquried in robotics. There are a various ways to consider this:
1) Google is being forthwright in its intentions to develop technology for industrial and manufacturing applications. Google has mapping technology, it has cloud technology, and many other capabilities. If you combine some of these into a robot you could, for example, have a robot that could do mining operations, still one of the most dangerous occupations around. There are indications the mining industry in general would like to get to fully autonomous operations. The industry is a signficant user of technology, such as location technolgies to route giant machinery in large mines.
Although manufacturing was stated, that makes less sense to me on the surface becuase there are so many large companies already deeply involved in manufacturing automation. The only way that makes sense is if Google believes they could deploy more general purpose robots that are easily trained and can do more human-like tasks (like Baxter 10X (a reference to Google's 10X approach.)
2) Google has decided that there is a close relationship between the internet of everything, robotics, cloud computing, and search-based intellegence and has the intent to grow a business in this area. Think IBM Watson. They started with a demonstration technology, now they want to leverage that into huge markets. The question is, what markets is Google really after?
3) Ths sinister view is that Google has gotten cozy with the Government, and is going to start generating revenues in a Government/Military business, as another way to leverage all the technologies they offer. Think about this--the large robots from Boston Dynamics connected to a massive cloud computing resource giving the robots near autonomy in dangerous areas. The earlier remark about Terminator is not far from the mark in this scenario. There are even more sinister lines of thought--as noted in another comment, let's say Google takes over mail and package delivery by robots. But these robots are linked to a massive cloud resources matching up terabytes of information relating to the shipping origin, the destination address, past delivery history, past history of all, say, catalogs and bills received, plus, it has sensors to sniff out chemicals, drugs, dangerous materials, etc., it can scan every package for all fingerprints and do further matching and connecting the dots. When you come to the door to get your robo delivery, it sniffs you and your home, adding further intel on possible drug use or other activities, perhaps it listens and makes note of how many people it thinks are there, what TV is being watched, what music is played, and even what's for dinner. Oh, and it also scans any faces it sees and matches them to a real-time cloud dB of people and flags if you might be somebody on a "list". The government then uses all this hyper-connected intel to "find" suspected terrorists, drug dealers, or other "persons of interest".
Which, if any is the right interpretation? Who knows.
eafpres, thanks for your detailed and thoughtful comments. I agree with you about the manufacturing application not making a lot of sense competitively. Unless, of course, there's some incredible innovation that will wipe that all out and leave it in the dust. The fact that BD was a pioneer in natural movements makes me wonder about that possibility: current factory robotic tech is still pretty clumsy and awkward. Some of the pick and place machines are very, very fast but that's only with a couple movements repeated over and over. So yes, I also thought of a revved-up Baxter. And I also thought about your #3--the sinister view. That's the one I think has people getting nervous. Good point about combining that with mail delivery and location (etc) information--yikes!
Even the most benign intentions can fall into the wrong hands. I agree with one comment in that Boston Dynamics has something GOOGLE feels is important to their overall goals and long-term vision. Boston has demonstrated their ability with robotic systems as Ann has demonstrated in her posts over the past several weeks and months. THEY ARE REALLY GOOD AT WHAT THEY DO. The fact that their largest customer is the "FED" really makes me nervous. (Please see NSA and millions of data points collected on a daily basis.) We are years away from "The Rise of the Machines" but significant intrusion into the daily lives of millions is happening right now. Boston Dynamics, GOOGLE, the FED. To me this could become an un-holy trinity. (Of course I've always been a little paranoid. Just ask my granddaughters.) I use robotic systems in my work but those are involved with automating manufacturing processes and developing work cells to provide added quality to the end result--the assembly of components. The accumulation of data from these work cells is used to calculate Six Sigma information, investigate trends and determine CpK. We DO NOT interfer with lives. Great post Ann and I feel your concern is right on.
I think the view in the article is a bit paranoid. On an overall level there really isn't any difference between commercial and military robotics. It is just the application that is different. The technical advances made by BD as a military contractor will be very useful in commercial applications. Nothing different from what's been happening for hundreds of years. Just think about how often all sorts of technology developed for the military has found commercial applications. To give an example, think of the work done by BD to make a stable platform that can find it's way around obstacles while carrying a heavy load. Now think about a robot stocking grocery shelves without trampling customers. Same robot, different application.
Until someone comes up with a battery that will last all day, I don't see Big Dog wandering around NYC or Boston etc. deliverying mail all day. Ditto, restocking shelves in the local WalMart store. The re-stocking could be done with today's technology in guided vehicles combined with a robot - at a horrendus cost.
Search and rescue, bomb disposal - pretty much already covered. And from a military standpoint this former Marine wonders how useful they would be when encombered with armor plate to keep the bad guys from blowing them away with a single 7.62 mm round. The American indian quickly learned that the US Calvery didn't have any armor on their horses.
I agree with RBPrice, although not just for battery reasons. BigDog delivering mail or stocking shelves seems like expensive, massive overkill. Not exactly an appropriate application of technology to the tasks at hand.
As the 3D printing and overall additive manufacturing ecosystem grows, standards and guidelines from standards bodies and government organizations are increasing. Multiple players with multiple needs are also driving the role of 3DP and AM as enabling technologies for distributed manufacturing.
A growing though not-so-obvious role for 3D printing, 4D printing, and overall additive manufacturing is their use in fabricating new materials and enabling new or improved manufacturing and assembly processes. Individual engineers, OEMs, university labs, and others are reinventing the technology to suit their own needs.
For vehicles to meet the 2025 Corporate Average Fuel Economy (CAFE) standards, three things must happen: customers must look beyond the data sheet and engage materials supplier earlier, and new integrated multi-materials are needed to make step-change improvements.
3D printing, 4D printing, and various types of additive manufacturing (AM) will get even bigger in 2015. We're not talking about consumer use, which gets most of the attention, but processes and technologies that will affect how design engineers design products and how manufacturing engineers make them. For now, the biggest industries are still aerospace and medical, while automotive and architecture continue to grow.
More and more -- that's what we'll see from plastics and composites in 2015, more types of plastics and more ways they can be used. Two of the fastest-growing uses will be automotive parts, plus medical implants and devices. New types of plastics will include biodegradable materials, plastics that can be easily recycled, and some that do both.
Focus on Fundamentals consists of 45-minute on-line classes that cover a host of technologies. You learn without leaving the comfort of your desk. All classes are taught by subject-matter experts and all are archived. So if you can't attend live, attend at your convenience.