Robotics Research Lab Willow Garage Shutting Down?

New submitter moglito writes "Willow Garage is/was acknowledge by many to be one of the best places for robotics research these days. Besides developing the PR2 it made itself a name for creating the open-source Robot Operating System ROS. But now it seems to be shutting down. [From a posting on the Willow Garage site:] 'Scott Hassan, founder of both Willow Garage and Suitable Technologies, said, 'I am excited to bring together the teams of Willow Garage and Suitable Technologies to provide the most advanced remote presence technology to people around the world.' Willow Garage will continue to support customers of its PR2 personal robotics platform and sell its remaining stock of PR2 systems. Interest in PR2 systems or support should continue to be directed to Willow Garage through its portal at www.willowgarage.com.'"

Comments

seems the pivot didn't work

[Trepidity]

Back in February, IEEE Spectrum reported that Willow Garage was shutting down, which led to a rebuttal from WG in which they said that they were changing, not shutting down. I guess the change wasn't profitable enough.

Really, Slashdot?

[Anonymous Coward]

From February of 2013

"Know it is shutting down"

[Anonymous Coward]

> But know it seems to be shutting down.

Why thanks! Now we know! And when, exactly, is it appearing to shut down? Now, you say? Why didn't you add *that* useful bit of information in the summary, editors? ... editors? ...

*chirp chirp*

Re:"Know it is shutting down"

[mrclisdue]

Well, bully for you for picking up on that,

However, six words in, we have this:

"Willow Garage is/was acknowledge

which you, apparently, missed.

So whom should we pooh-pooh for lack of editorial skills?

cheers,

Very sad

[Anonymous Coward]

Willow will be missed. After so many contributions to open source and robotics, it's absence will definitely be notisable.

ros.org

[Anonymous Coward]

After this announcement there was a new release of ros with promises for another in 2014, so that looks to be going strong. That's a relief.

New home for ROS

[Anonymous Coward]

The OSU Open Source Lab is picking up the Robot Operating System's web presence (ROS being the software that powers WillowGarage's units). It's been a long process, as they have a lot of moving pieces that we're integrating, but hopefully the entire setup will be completed next week and we'll have an announcement to make.

manufacturing problem, not software problem

[stenvar]

The problem with robotics and its failure to catch on widely I think is largely related to the fact that robots are still expensive to manufacture. Willow Garage doesn't seem to have made much progress with that. If you could put the hardware for an arm or a human-height telepresence robot in people's hands for less than $1000, the software would take care of itself. Working on 'robot operating systems" and similar software right now probably remains wasted effort; by the time the costs for the hardware has come down, all that work will likely be obsolete anyway.

Re:manufacturing problem, not software problem

[Spiked_Three]

You may be right. But you are over simplifying it. Robot magazine (or was it servo?) published a series of articles where someone basically built their own robot compatible with PR2. I think it was in the $1000 price range.

It's more like even something like the $500,000 pr2 is still pretty useless for anything but research. Prices need to come down orders of magnitudes more. Chicken or the egg? Until someone can market a $5 million dollar robot for $299, human size, robots will never make it into mainstream.

Frankly our only hope is that such a thing is developed for war.

Re:manufacturing problem, not software problem

[stenvar]

If you can build it yourself for $1000, it probably would cost $10000 as a product; just look at 3D printers.

I think two arms on wheels would be extremely useful, and if you could sell it for $1000, people would snap them up and solve all sorts of robotics problems in a heartbeat.

The PR2 itself is ugly and overengineered, though; it's the kind of monstrosity academic groups with too much money have been producing for decades. That's not something anybody would want to have at home, no matter how useful it may be.

Re:manufacturing problem, not software problem

[Anonymous Coward]

The expense can be large depending on what a robot is required to do, but the hardware for a $1-2k telepresence robot is certainly out there today. It's as easy as sitting a monitor on top of one of Willow Garage's turtlebots. A motivated hobbyist could probably accomplish the same with a raspberry pi, webcam, montior, and some R/C car parts for much less.

The hard part is, once you've built your robot, how to program it to accomplish even the most basic tasks. That part most certainly does not just "take care of itself" as you describe. How do you write code that tells the telepresence robot to "meet me in the kitchen?" How does it know the difference between a black rug beneath it and a cliff (my roomba certainly doesn't.) Companies and amateurs alike had to either resort to using some sort of robot framework software that didn't quite do everything they needed, or rolling their own software complete with re-implementations of navigation, localization, mapping and other algorithms. These frameworks also have to handle transporting sensor data around in a usable format to all algorithms that needed it, handling actuator commands, and generally merging different data sources together to enable advanced functionality.

What ROS, and all of the other frameworks that came before it, are trying to do is to create a standard library of autonomous behaviors, and a well-defined way to communicate data and commands between producers and consumers. The difference now is that ROS is starting to become the de-facto standard among academics, and it's even moving out into industry, something none of the previous frameworks have achieved before. Indeed, a lot of the other frameworks that are still alive and kicking are working towards becoming ROS-compatible so as to stay relevant. Standardizing on a common platform means you spend a lot less time trying to work out the plumbing, and more time working on the problems you actually want to try and solve. Only in that way will the harder problems start to 'take care of themselves.' Hardly wasted effort.

Re:manufacturing problem, not software problem

[Spiked_Three]

No, the problem is people that think they understand, when they do not.

First, a telepresence robot you describe is really a remote control camera, not a robot. But yes, a dime a dozen.

Where did you get the idea to use the word behavior? It has absolutely nothing to with what you are describing.

Difference between a black rug and cliff. I don't know maybe a distance sensor for christ sakes. again, stop guessing. The fact that your roomba didn't add a $.35 sonar sensor IS a manufacturing problem just as described. The fact that roomba couldn't be bothered with slam, like Neato did, is a consumer rip-off problem.

You are semi correct that ROS was intended to create libraries (not behaviors) of common useful tasks. Totally wrong as to previous frameworks. But as usual, and here is where I will get into trouble with the open source crowd again, the open source implementations were/is horrible. Microsoft robotics, eventually lead to WCF, and is even being built into the next generation of hardware (multiple processing units attached to multiple storage/sensing units on common bus). ROS lead to ... nothing. Maybe someone with some vision will wake up at Microsoft and identify a commercial venue and re-pursue robotics, but until someone with profit as a motive gets involved, robotics is doomed. Will Google save us? Maybe if Apple sells a gold colored robot people will sleep in line for it.

Re:manufacturing problem, not software problem

[tlhIngan]

The problem with robotics and its failure to catch on widely I think is largely related to the fact that robots are still expensive to manufacture. Willow Garage doesn't seem to have made much progress with that. If you could put the hardware for an arm or a human-height telepresence robot in people's hands for less than $1000, the software would take care of itself. Working on 'robot operating systems" and similar software right now probably remains wasted effort; by the time the costs for the hardware has come down, all that work will likely be obsolete anyway.

The big problem with robotics is its a marriage of two very distinct engineering disciplines - mechanical engineering, and electrical. And it's not easily separable, either - both disciplines have to work very closely together because it's easy to screw yourself either way, and the two disciplines don't generally interact this closely.

Take for example a robot joint. Now, the mechanical engineer might want to simplify things and eliminate costly gears or a tricky mechanical mount, but in doing so, it puts demands on the electrical side - perhaps higher torque motor, or one that is more precise. Or maybe the feedback devices need more precision - if it's a simple pot, perhaps you need higher quality power supplies (less noise), coax cabling because you now need a more expensive 20-bit ADC where before you could get away with a cheap 16 bit one. These are huge tradeoffs that have to happen - perhaps wanting to use cheaper electronics results in more complex mechanical systems that may cost more to fabricate, or lower reliability.

It's why universities have a middle of the road option called mechatronics which combines both at a high level, because it's not an easy problem to solve - where do you draw the line? And experience is needed - when can you get away with cheaper parts, and slightly more complex mechanicals, or is it better to spend more on the electronics and less on the mechanicals to get a reliable system?

And what about failsafes? If the motor is full of torque, what happens when it encounters resistance? Do you push it away, maybe burning out or ruining the mechanics or blowing the fuse? What if the resistance was a misplaced hand? What if someone pushed the E-stop? IF it's a hand, yes, you want it to release pressure. But what if it was holding up a heavy engine block? Releasing power could hurt somebody when it moves, indeed, it could pin someone else down. (Generally speaking, robots and humans are kept separate to ensure that E-stop would simply de-power equipment and as long as no one is in the danger zone (which they shouldn't be if the robots were on), it doesn't matter.)

Robotics is not a simple task, and it's not just manufacture, but also engineering and R&D. Manufacture is just one aspect.

And yes, once the hardware is figured out, the software can be concentrated on later - each hardware decision has a ripple effect to the software.

Re:manufacturing problem, not software problem

[RespekMyAthorati]

something like the $500,000 pr2

At $50,000 the PR2 would be grossly overpriced for what you get.

WTF?

[Spiked_Three]

So Microsoft abandoned robotics.

Open source (Willow Garage) is abandoning robotics.

Who is going to fulfill my lifelong fantasy started as a child of robots in real life? Google?

I do not understand how something so pivotal in humanities history, can become so hopeless.

Then again, I remember how completely let down and disgusted I was when 2001 came. It was nothing like the movie.

Turns out I have had false hope about mankind all along.

ROS

[hammeraxe]

ROS is a really good framework. And the best thing is that it's open-source so even if WG will go down ROS will continue to exist

Re:ROS

[Anonymous Coward]

I'm working with some former Willow Garage employees to ensure ROS lives. It is by no means trivial to keep ROS alive, with their large set of jenkins jobs, their insanely massive set of documentation, or the large set of ROS packages themselves. I believe my mirrors will be their first non WillowGarage set of source / package mirrors, and I doubt anyone just happens to have the entire source archive sitting on their homedir, so we should keep in mind that "open source" doesn't mean "lives forever" at the scale of Linux Distrbutions.

Shutting down vs spin-offs

[movrev]

Willow Garage is not so much shutting down, but rather being dismembered. You have Suitable Technologies, Industrial Perception, Open Perception, the Open Source Robotics Foundation, and Redwood Robotics as spin-offs (and I must be forgetting some). In my opinion, it never had a very good business plan in place (given the cost of the PR2 and its capabilities), but it does have the right open source attitude, which helped cement ROS as a standard in the robotics community.

That's too bad

[Animats]

That's too bad. I've met many of their people. They were doing good work.

"Suitable Technologies" is just another company producing those annoying "remote presence" robots with a video phone on top. There are four or five manufacturers of those things. They can't do anything; they have no manipulation capability. They just talk.

"Remote presence" is only useful if the person running it is someone who gets sucked up to. Like doctors. A number of health care vendors are trying the things.

12 hours, 16 comments

[Spiked_Three]

That pretty much explains it, right there.

Miley's tit size would generate more discussion on slash dot.

Why is robotics so ignored/boring/avoided, by even a tech community?

I am seriously starting to believe what I said earlier, about war being robotics only chance. Fly a drone remotely, kill people, bam, tons of interests. Let's add walking robots with fully automatics to it and invade, I don't know, pick any middle east country, maybe iraq again. Sorry folks need to die, but progress must be helped along, so we can use robotics to help people.

Re:12 hours, 16 comments

[mpfife]

Why is robotics so ignored/boring/avoided, by even a tech community?

It's a hardware problem. Most people here are software people. :D

Re:12 hours, 16 comments

[RespekMyAthorati]

Actually, guys at CM, MIT and Stanford (and many other places) are steaming ahead with robotics.
And don't get me started on Japan.
Another place is at the companies who are trying to build a self-driving car.

Old news is old

[hirschma]

That story was from the 21st. Of August :) If a robotics company falls in the Valley...