Microsoft Developing Robotics Software

s31523 writes to tell us Microsoft recently announced the launch of their new Microsoft Robotics Group and the first product release, a software program to help robotics developers. Despite the timing this has nothing to do with the recent abdication by Gates, and was actually instigated by Gates before his departure. From the article "It might take many years, but Microsoft believes robotics could present a big opportunity as the market grows, said Tandy Trower, general manager of the Microsoft Robotics Group. He cited estimates predicting that consumer robotics alone will grow into a multibillion-dollar industry in five to 10 years."

1st BSOD?

[MMHere]

What happens with the first BSOD. Will the robot fail to avoid Asimov's First Law if in motion at the time?

Crazy tangent?

[Otter]

And no, despite the timing, it's not a case of the company's engineers taking Microsoft on a crazy tangent now that Bill Gates is shifting away from his day-to-day oversight.

What "crazy tangent"? Every robotic system I've ever worked with was controlled by software running on Windows (or DOS).

Trouble branching out

[LehiNephi]

Microsoft has been experiencing for several years what Google is only now realizing: They're good at a few things, and that's it. Microsoft, feeling the pinch of having essentially only two major products (Windows and Office), wanted to diversify. While they have a near-monopoly on operating systems and office suites, that's the only market in which they have a large, profitable stake. So they try to branch out. Sometimes, they're more successful, like with their mouses. Sometimes, they're not, if you look at the financials of the Xbox. The problem they face, however, is that the markets they want to branch into are already well established. Crowded, even. So MS throws piles of money at it, hoping that it will work. At the same time, Apple and Linux are starting to make inroads in the desktop and server markets. MS sees their mainstay threatened.

Google is similar. They came up with a great product, their search engine. It was so good that it rapidly took a majority of the market, despite default IE settings. But then they stalled. GMail is good, but has nowhere near the market penetration as their search. Maps, groups, IM, blogs, calendar, spreadsheets...the list goes on. Google has some good products, but they're trying to expand into an already saturated market. And now their flagship product is faltering. Linkfarms, SEOs illegitimately boosting their rankings, and spammers are degrading the quality of Google's results.

Now, we're not talking about a mature industry with human-interactive robots. However, this smells strongly of "We need to find a new way to make money if Windows/Office starts slipping"

OMG

[infosec_spaz]

The END in Near! I can see it now...Robots running amok, pulling the heads off of small furry creatures, killing babies, mameing everything in there path, then, BSOD.

Re:Real time Windows?

[Otter]

I'm prepared to accept that there are plenty of programming systems for automation that are Windows-based, but actual robots?

Most "robots" *are* just automated devices, not Commander Data-like sentient androids. Take this Quadra 3 SPE [tomtec.com], for example. (Note: Windows-based!)

That's why I always laugh when people here spout off about Asimov's Laws in connection with industrial robotics. It's like complaining that your toaster oven should know not to burn you.

why so much fuss?

[Kalinago]

Microsoft is definitely a newcomer in the territory of designing applications for this market.

On the other hand, most industrial robots for Welding/Automotive/Manufacture production are basically soulless drones that follow repetitive sequences of greater/lesser complexity written in ladder logic or some proprietary language; and the "brains" is generally a PLC. Popular proprietary PLC systems (Rockwell, Siemens) rely on Windows based software to download your ladder logic program and update the firmware. So it's still Windows after all.

So in theory, this is a market where microsoft should not encounter much trouble.

I believe most /.'s are concerned over high end robotics programming; truly making decisions, neural network based, AI, vision controlled ones, path finding...'top of the heap' applications that are non mainstream and limited to research or hi-tech chemical/petroleum/aerospace industries. I don't know much about numbers, but I doubt that this is makes up a significant market share, even today. So I'm betting Microsoft eyes may be set upon the first option, as most plant floor operations are becoming fully automated even in developing countries.

--
forget past mistakes, and condemn yourself to repeat them.

Uh huh

[ObjetDart]

He cited estimates predicting that consumer robotics alone will grow into a multibillion-dollar industry in five to 10 years."

I remember the last time I heard that, it was... oh, about five or 10 years ago.

Microsoft Robotics? Are you crazy?

[plusser]

When it comes to robotics, Microsoft need to understand that they are not electromechnical engineers. There have been many "False Dawns" with the idea of robitics in the home, many problems are down to the fact that the robots need to interact with the most illogical lifeform on the planet - Man! When you consider that the market leaders in robots are mainly Japanise Car Manufacturers, whom only build demonstration models to show off how good they are at building robots. I can think of only two companies that have attempted to sell robots in the domestic market, Sony and its err.. Dog, and Dyson with a robotic vacumm cleaner.

The biggest problem with robots in our homes is safety. No only does the robot have to perform complex tasks that may appear easy to humans, but it also has to ensure that humans do not come into danger as a result. With the kind of blame culture in the West, it would be crazy to think that anybody will enter this market without understanding the implications of a lawsuit. That's why robots are good in environments where human access is restricted, such as the factory or on a space mission.

My advice to Microsoft is simple, continue what you are good at - screwing all those companies (especially those with less ethical business practices) with your high priced Operating Systems and Office Solutions for use in business IT systems. Yes, those of us in the know will continually priase Linux or Apple (and save lots of money in the process by buying a more suited product) and maybe think that the XBox is possibly a good product.

However, if Microsoft think they can bring some innovative to the market, they better get in contact with the high reliability electronics market - robots are not going to be consumer devices anyday soon...

5 to 10 years?

[Pedrito]

He cited estimates predicting that consumer robotics alone will grow into a multibillion-dollar industry in five to 10 years.

The guy who cited these statistics probably agreed that 640K was more memory than anyone would ever need.

First of all, there's simply nothing to base this on. How many households currently have consumer robotics? Percentage-wise, it may as well be 0%, because it's pretty damn close to that. So how can you possibly predict that consumers are going to buy billions of dollars worth of something that doesn't even exist in anything other than a manufacturing, hobbiest, or neat but useless gadget category?

Before you can make a prediction like this, we really ought to see one or two robots that look like they might do something consumers would want. And don't even tell me about the robotic lawn mowers. Show me one that doesn't involve border wires (most people don't want to be bothered) and doesn't have to be monitored so it doesn't run over the dog/cat/baby. Robotic vaccuum cleaners, maybe, but show me one that has enough power to really vaccuum, isn't bound by a cable, and can navigate a staircase.

Sorry, but I simply don't believe we're 5 to 10 years away from robotics being a "multibillion dollar industry". 15, maybe 20, but not 5-10. I just don't see it happening. Robotics simply hasn't progressed all that far in the past 10 years compared to a lot of other consumer electronics (DVRs, computers, iPods, etc)

Re:Crazy tangent?

[Otter]

I'm a roboticist and I have to take issue with this sample bias.

Sure, I'm not making assertions about market share at all. (I don't have the slightest idea, and you and some others replying clearly do.) I'm just saying that the use of Windows in robotics is hardly as unprecedented as the link makes it sound.

Sooo, this was how BORG came into being eh ?

[unity100]

They didnt give info on its history in Star Trek. It seems they didnt want to spoil the fun : we are going to see what happens by LIVING it.

Re:Crazy tangent?

[ScottLindner]

VxWorks is only used in complete system designs. Things that NASA would do are certainly true for this case.

But other robots use controllers like PLCs, CNC, and another hybrid that I can't quite remember at the moment. These are much simpler and are not what I'd call an "engineered" solution but a very simplistic one. It does depend on what the robot is. Which is a very fuzzy distinction without any clarification. Which I guess is the basis of my point.

Vx didn't work for a long time... (a joke with some truth)

Bill Gates is quite aware of the Singularity

[vertinox]

And I quote... Off the back cover of the book, The Singularity is Near [wikipedia.org] by Ray Kurzweil [wikipedia.org]

"Ray Kurzweil is the best person I know at predicting the future of artificial intelligence. His intriguing new book envisions a future in which information technologies have advanced so far and fast that they enable humanity to transcend its biological limitations--transforming our lives in ways we can't yet imagine."
-Bill Gates

This single quote has made me go "Hrm... Thats odd." If you are a futurist/transhumanist advocate it is understandable why you would advocate the book. However, a straight faced businessman who happens to be one of the most wealthiest men on the planet (next to that guy from ikea) starts to laude and praise this book at the future... Well... It makes me wonder what Gates has planned.

If you haven't read this book, then get it, put some time aside and give it a thorough reading. I'm sure there are something things that we all disagree with in the book (including myself) but it has to be one of the most logical explanations of the Age of GNR (Genetics, Nantotechnology, and Robotics) we are about to embark in 10-40 years.

With that in mind, I believe Robotics is the next big boom (as the internet was in the late 1990's) and within the next 10 years robotics will have affected us more than internet has. Think Roomba, DARPA Urban grand challenge, unnamed flight, and so on...

I wouldn't put it past Bill to know what is going on here (although he did bungle on predicting the importance of the internet back in the early 90's).

So I think this is an attempt to at least be in the game if and when the robotics boom arrives.

Re:Microsoft Robotics? Are you crazy?

[ThinkingInBinary]

I can think of only two companies that have attempted to sell robots in the domestic market, Sony and its err.. Dog, and Dyson with a robotic vacumm cleaner.

Um, iRobot [irobot.com]? The Roomba [irobot.com] (and the Scooba [irobot.com], more recently)? They're much more well-known than Dyson's robotic vacuum, and much more useful and reasonably priced than the Aibo. The Roomba and Scooba seem to fulfill the goals of safety and functionality. They have a simple enough task that safety is just a matter of stopping if something is in their way, and they have dedicated hardware to do their job.

Admittedly, they aren't general-purpose robots, but I predict that specialized robots like the ones iRobot makes will be much more commonplace and useful than humanoid robots that can vaguely interact with real humans, wander around, and not do much else. Building a robot for the purpose of vacuuming or mopping a floor is 1000x easier than building a robot that could learn how to vacuum, if you gave it an upright vacuum and it had the dexterity to operate it. Robots should not use human tools, they should be tools.

Re:1st BSOD?

[LordEd]

More like: "It looks like you are commiting a crime. Do you want help with this feature?"

Re:Crazy tangent?

[ScottLindner]

I'm not sure it's as bad as you think. The term "robot" is such a vague concept that many things could be called a robot. And they are all correct. For each design and implementation, the creator chooses what is best for them based on current skills and how well those skills match the need. Sometimes this means rolling your own, sometimes it doesn't. It depends in every case. Much like a car. I'm sure you know that the computer in one car is not interchangeable with a computer in another car. They are preprietary systems for good reasons for each maker of cars. This is true for robot makers as well. But going even further, what you consider a robot also greatly influences the divesity of robot control products.

I built robots using PLCs as a kid. It's very simple to problem and the programming language/model matches the design of the robot itself. hard to explain unless you know ladder logic (relay logic). For a very long time this was the *only* way automation was done in the industry. It still is used very heavily because it doesn't require a great deal of sophistication, education, and is not an engineered application for the task. PLCs are intended for electricians to program on the factor floor. But the complexity of the program is very limited.

You have other robots that require something much more sophisticated to control them that design of is just as integral to the end robot solution as the physical robot itself. This is probably where you see a lot of VxWorks being used.

I'd suspect Windows applications are being used as a replacement for the PLC type of applications.. but to enable something more sophisticated because it's easier to write software for a complex situation than you could do with ladder logic in a PLC. But in this situation, it takes engineering support to automate, and not a floor technician.

This is a very simplistic way to address your concern. Think of NASA's robots. You think they should plunk an ATX mobo and CPU in there with windows on it... or build their own RAD hard, high G, extreme temperature tolerant, liquid cooled only, lightweight computers that are a trade off of weight, survivability, and just enough computing power, or should they use some standardized solution to reduce the number of skills required? Depends on the needs.

Re:Trouble branching out

[notaprguy]

Who voted this "insightful?" Too many silly statements to summarize so I'll just pick on one of the most obvious.

Nephi said "While they have a near-monopoly on operating systems and office suites, that's the only market in which they have a large, profitable stake."

So...what about SQL Server or Exchange Server? SQL Server alone would be one of the biggest software companies in the world and Exchange isn't super far behind. Both are highly profitable.

I also can't resist tweaking the equally silly suggestion that it is somehow new that the markets MSFT is entering are "already well established"..."crowded even." I have news for you. Almost every market MSFT has ever entered was already established. Spreadsheets? Lotus 1-2-3 was there first. Word processors? There were many before Microsoft Word. There is nothing new about Microsoft entering markets that are already established. That is, in many ways, their specialty: finding profitable markets that others pioneered and competing with decent (but not always the best) products and good (but not always the lowest) prices.

Nuh... don't need them

[rajafarian]

Microsoft will go through the motions, make announcements, hire some people, make more announcements, show a demo, make more announcements... but when it comes time to compete, they will do the only thing they know how to do:

Buy off a company who actually has something to show for their efforts!"

Re:1st BSOD?

[colmore]

It's _Asimov_ and, well, duh.

Every single story he wrote that used the three laws was about how the three laws couldn't really work.

So quit being such a frikkin' know-it-all, you humorless chump.