Robot Assembles Ikea Furniture
nytimes.com
nytimes.com
It also explains why no robot is nearly as agile as a mouse, while computers routinely beat human world champions at many ‘complex’ games.
"Moravec's paradox is the discovery by artificial intelligence and robotics researchers that, contrary to traditional assumptions, high-level reasoning requires very little computation, but low-level sensorimotor skills require enormous computational resources. The principle was articulated by Hans Moravec, Rodney Brooks, Marvin Minsky and others in the 1980s. As Moravec writes, "it is comparatively easy to make computers exhibit adult level performance on intelligence tests or playing checkers, and difficult or impossible to give them the skills of a one-year-old when it comes to perception and mobility".
Similarly, Minsky emphasized that the most difficult human skills to reverse engineer are those that are unconscious. "In general, we're least aware of what our minds do best", he wrote, and added "we're more aware of simple processes that don't work well than of complex ones that work flawlessly"."
"As Moravec writes:
Encoded in the large, highly evolved sensory and motor portions of the human brain is a billion years of experience about the nature of the world and how to survive in it. The deliberate process we call reasoning is, I believe, the thinnest veneer of human thought, effective only because it is supported by this much older and much more powerful, though usually unconscious, sensorimotor knowledge. We are all prodigious olympians in perceptual and motor areas, so good that we make the difficult look easy. Abstract thought, though, is a new trick, perhaps less than 100 thousand years old. We have not yet mastered it. It is not all that intrinsically difficult; it just seems so when we do it."
you know how to party!
Does make me wonder where programming fits into this. It seems to be both hard for people and hard for AI.
Look at the Cyc project. After decades of work, what have they got?
So I push back on the idea that computers have already been taught what is essentially REASONING.
Why isn't teleoperation a more heavily researched area? I suspect that it could be efficiently accomplished with manipulators that are much simpler than human beings or human hands, so long as you add two powerful ingredients:
1) Movable point of view. Basically this could be slaved to the position of a high resolution VR headset. It could even be provided by a tethered drone.
We've had teleoperation for a long time, but it has generally been hampered by a static point of view. What if you could crane your neck and look around what you are doing? This greatly improves the intuitiveness and effective bandwidth of the interface.
2) Haptic feedback.
If you can feel what's happening, even indirectly, it also greatly improves the quality of the interface. Think of the dexterity you have with a pair of forceps or needle nose pliers, versus the dexterity you'd have with those while your hands are numbed by novacaine. Since such mechanisms would be much simpler than a mimic human arm, they'd be much cheaper. Also, human beings are very adaptable to control tasks and don't really require a precise mimic of human limbs. (Witness the amazing feats of backhoe control you can find on YouTube.)
Haptics can greatly degrade with lag. There have been some approaches to predict what the operator is going to do or predict that a collision is going to occur and prevent such things, but they have not been able to decrease the time it takes to complete a task[0].
It is worth noting though that remote surgery has been somewhat successful over high bandwidth communication links. One constraining factor for teleoperation research is that there aren't many applications where it makes sense to do so. The problem is that robots and haptic gear are currently expensive which limits the applications. You need an application where humans cannot go or the labor the operator is doing is more valuable than the cost of transporting the operator there and the robot. Applications where humans can't go are fairly niche. In addition, applications like working inside a non-damaged nuclear reactor or subsea operations tend to involve equipment that has been designed to be serviced by the simplest of teleoperated robots. In the case of nuclear reactors, mechanical and hydraulic remote manipulators are suitable for most operations. Very valuable tasks like robot surgery are rarely done remotely, partially because of how expensive surgical robots are. We could put surgical robots in remote locations, but remote locations have less patients to service, so it's hard to justify the cost of putting robots in a bunch of remote locations versus moving the patients to the hospitals.
[0]https://ieeexplore.ieee.org/document/7399376/ [1]https://en.wikipedia.org/wiki/Remote_surgery
JAXA has been looking into this. It seems that highly motivated operators can tolerate a full second of lag.
In the DARPA robotics challenge, competitors were able to create movable views with lag by 3d scanning the environment and streaming the 3d scans.
That's entirely suitable for a "work bench." In fact, the background might as well be blanked out and replaced with a synthetic grid.
You need an application where humans cannot go or the labor the operator is doing is more valuable than the cost of transporting the operator there and the robot.
I can envision many contexts in the upcoming decades where teleoperation is going to be very useful. The long term health costs of human beings suiting up to go into environments without protection from radiation could be quite high.
Now before we can see this approach used in practical applications, they are going to have to improve the reliability of assembling things. If the reliability of manipulating things can be improved, this could be very disruptive. I think that near term, the killer app for this is an autonomous robotic welding cell. In 2012 it was shown that welding work for a vehicle frame could be autonomously planned[0]. Robot manipulation was not good enough at the time, so the robot had to tell the human where to put all the fixtures and pieces to weld and then welded them. The human also had to reorient the parts too. The implication of this work, is that many of the operations they used to make the chair are also applicable to autonomous welding. This could be very disruptive as it could make it practical for robots to do welding for low manufacturing run parts. Low manufacturing run parts tend to be made by human welders now.
Full assembly 8min 55sec.
(I'm thinking of something like Bender with an accent, but hey that's just me.)
And if it was that hard you wouldn't see so many people with so much Ikea furniture out there. :)
The robot was "fed a kind of manual, a set of ordered instructions on how the pieces fit together" - if all it was given was pictures of the real manual, that would be quite cool, but I'd love to see more details here.
But of course IKEA doesn't want to pay translators to write that sentence in 20+ languages, so pictures only it goes.
Label the pieces by type: A, B, C, etc
Add corresponding annotations to the manual.
Other brands do this. It vastly reduces the amount of time spent "is this the right piece of plywood? Or is it this other one? Which of these two screws?"
Some context through text would help even more in certain places, but you don't even have to go that far to make it a lot easier for people not familiar with matching parts to diagrams.
What's fun about that nightstand is that you have to really look at the picture in details to figure out which side up and which side forwards for the drawer slider stuff.
Also tou need to assemble everything laid out as it's displayed in instructions without any improvisations. It's not the case of "if everything else fails RTFM"
Mayonnaise is the most popular sauce ever; yet the meme is that it's "disgusting".
Almost everybody has Ikea at home; yet the meme is that it's impossible to assemble.
Maybe memes are trying to tell us something... but I'm not sure what.
We should take this back to the original formulation of "meme" by Richard Dawkins.
https://en.wikipedia.org/wiki/Meme#Origins
Maybe it's the opposite.
Memes, in the academic sense, are just ideas that can propagate very well. Ideas don't have to be true to propagate.
Mayonnaise is the most popular sauce ever; yet the meme is that it's "disgusting".
Disgust is a very useful emotion to aid in the propagation of an idea. Outrage is another.
https://www.youtube.com/watch?v=rE3j_RHkqJc
https://hbr.org/2016/05/research-the-link-between-feeling-in...
I don't personally care for mayonnaise. I think the memes are popular precisely because mayonnaise is so popular. A lot of the anti-mayo memes I've seen complaining about mayonnaise are about finding it where you don't expect it or ordering sandwiches without it and still getting it. There are fewer memes ranting about less common ingredients like Buffalo sauce or hummus because you don't have to go out of your way to avoid those.
Just spit balling here:
1. Exaggerations can be funny
2. We, as a society, have a general tendency to go with the flow when it comes to trivial overstatements, even if we don't agree with them ourselves (maybe in an effort to "fit in")
3. If you hear the same lie enough times, you might believe it
4. Picking on a particularly popular name brand makes a generalization stick in your mind better
5. Lazy writers rely on cliche to get their points across
That is probably not true of most people in the general population.
The point of the article is that it is non-trivial for humans to make robots able to do what they can do easily themselves.
The IKEA task was likely chosen because it is something many people have personal experience with, which makes for better public understanding and engagement with the underlying engineering work being displayed.
Totally off-topic but keeps bugging me everytime I stumble upon it: The decades old cliché that Ikea instructions are tough to understand and the assembly will always fail (at first try).
Is this some anti-globalist urban myth?
Has anyone claiming this ever looked at the assembly instruction of furniture of smaller companies? Every mid- to high-class designer furniture item I've bought came with instructions that had terrible UX, that were over-detailed, cramped onto one sheet of paper, had way too small print size to read, made it difficult to differentiate screw sizes etc. virtually all of them left you at one point in uncertainty how to interpret the next step, which could cause destruction if done wrong. Some even came with outright wrong instructions.
Contrast that not only to the text-free Ikea-pictograms that work across the globe, but also to the thought that went into optimizing the assembling so that complicated "points of no return" are virtually absent.
People think that because the visualizations are relatively simple in their shapes, then the design itself must have been simple. It's exceedingly difficult to create instructions that work across language and cultural barriers. It takes a tremendous amount of thought and effort, and IKEA is almost certainly the best in the world and deserve recognition for that.
"Simple ain't easy"
I think both myths comes from the same innocent place - that it's fun to take swings at a big guy who won't get hurt or swing back.
However, one Ikea product, the formidable “Liatorp” storage unit, which stands at 7-feet tall, has been dubbed “the divorce maker” by one marriage counselor who says the product’s 32-page instruction manual and 169 screws are enough to test even the most solid relationship.
Does said marriage counselor offer us any better "formidable 7-feet tall storage units" by competitors? No. Because for that, if the "divorce maker" claim has any merit, we'd have to ask the local precinct to find out about their experiences with other cabinets.
IKEA makes things that just about anyone can assemble, including people who would never try to build any other furniture. Those same people may be the ones propping up the myth because building IKEA furniture is harder than purchasing pre-built stuff from the store.
Admittedly, it's really easy for me to read/interpret the instructions. I could see how assembling furniture would be a challenge if you're not an engineer or something like that.
Next step: don't even do that. Forget you ever saw the thing in the showroom. Just take the pieces and and figure out what they make. Then afterward, compare what you built to photos in the catalog.
People have different kinds of capabilities, and IKEA furniture assembly may be easier for developer/engineer-types than others.
The components don't have labels and often require careful inspection to choose the right one, based on illustrations that are also unlabeled and low-resolution. Often times the pieces can be assembled incorrectly in a ways that don't become evident until later, requiring disassembly to correct the mistake. I can easily believe that people who have weaker visual-spacial skills and/or are less detail-oriented would find the assembly hard and frustrating.
Someone made a good analogy elsewhere: assembling Ikea furniture has a lot in common with some kinds of puzzles. Some people love puzzles, others tolerate them, and some hate them. That's not "just a meme."
It's not especially hard to do, it's just a meme to complain about it.
Because their products are cheap and relatively stylish, so people bear with it? IIRC, their success comes from the efficient logistics that this kind of design enables.
Edit: I forgot to add, I think everywhere IKEA also offers an assembly service, even same-day if you do it early enough. And with the assembly cost added, I expect their furniture is still cheaper than much of their competition locally.