Slashdot Mirror
"Augmented Reality" For the Assembly Line
silkySlim writes "EETimes has a short article about a combination data goggles and earpiece device to replace big manuals and reduce training time for assembly line workers. 'In one possible scenario, a technician with data goggles bends over the engine block of a luxury car and removes the covering. He is receiving instructions through an ear piece telling him what to do next while his data goggles mark the screws and bolts on which he must next place his tool.' Apparently, it's already in use by several automotive companies. There's some additional papers also available."
Tank, I need to know how to fix a Lexus IS300, now!!!
This post cannot be rebroadcast without the express written constent of Major League Baseball.
THE GOGGLES! THEY DO NOTHING!
Now if only these came with a bedroom module then maybe I'd know what to do when I found a girl in my bed.
I would be interested to see what in the way of technical writing and documentation goes into this sort of thing. It would force many technical writers to also focus on interface much more than they do now with standard Robohelp systems or other standard documentation.
Visit Jonesblog and say hello.
Until they start implanting happy thoughts and images into the system to keep the workers productive. "You love your job."
Talk to most anyone who works on an assembly line and they will tell you that their job is as boring as hell. Some might say that they almost feel like a machine themselves.
Contrast this with a labourer who builds furniture from scratch or a shoemaker and you find yourself in a different situation. While their actions are the same, their efforts have tangible results. If they have their own business selling what they've created the satisfaction runs deeper.
How are goggles shining lights in your face saying "Unscrew this next" going to make you feel any closer to the work that you're doing? It just seems like another level of detachment to me.
Why did the article discuss hardware problems but not social problems regarding the goggles?
Mind you, from a Slashdot geek perspective, the goggles are a cool idea, but I don't feel envious of the people who are going to have to use these things on an assembly line.
Sure its definately cool, but its scary too. Imagine directions being constantly spoken into your ear as you perform your job.
Human beings are not efficent organisms, neither in chemistry or psychology. You can't expect a human to act like a machine, something has got to give.
On Wall Street they say "buy low, sell high" On the pad we say, "buy high, sell high" Isn't that somehow better?
I tried one of the AR displays a few months back. The main reason why I thought that particular model was unsuitable for prolonged use, was that the text and other information appeared at a different depth from the object I was observing. The AR information was displayed at a fixed "infinite" depth.This made it impossible to focus on both the text and object at the same time, requiring me to adjust my eye focus everytime I wanted to read something. This constant refocussing caused a good amount of discomfort.
Adjusting the depth of the text to make sure it is exactly superimposed on the object that is being viewed is quite challenging, especially when the viewer moves his line of sight frequently.
An alternate design that some people find easier to adjust to uses a display mounted on a single eye (with the RW showing up in a dimmed background). I haven't tried these, but supposedly they are easier to get used to.
An Indian-American Hindu committed to non-violent thought/speech/action alarmed by the global explosion of radical Islam
The four biggest problems with extensive adoption of this idea are:
1. Safety and Liability. I can just imagine a bug telling the new assembly line "cyber drone" to drill a hole 1 foot to the left when it meant one inch (shades of Nigel Tuefnel!), and the resulting explosion when he drills into the fuel tank. That, and the possibility of anyone who screws up telling his supervisor "Hey, that's what the Magic Smart Goggles told me to do!"
2. Cost. Technical writers are comparatively cheap [and easy to lay off, he noted bitterly]. Programmers are expensive. If the new Mark 2 Framistan has holes in a different places, that's five minutes of work tops to put the new information in existing manuals, but a day to write the code, debug it, and test the magic googles to make sure they're acurately pointing out the new framistan holes rather than the old ones.
3. Limited Applicability to Modern Manufacturing. A good portion of the most repetative assembly line jobs have already moved overseas. Many of the mechanical assembly jobs left don't require one worker doing the same thing 100 times, but doing 100 different things on a far more complex tool (i.e., the difference between assembling a toaster and assembling, say, an Ion Implanter). Optimizing "Enhanced Reality" for one task performed 100 times a day may be cost effective, but not for programming and training the system for hundreds of tasks.
4. The Awesome Power of Human Stupidity. Everytime they make something idiot proof, nature has shown the amzing ability to come up with a better idiot.
Lawrence Person (lawrencepersonh@gmailh.com (remove all "h"s to mail)
http://www.lawrenceperson.com/
Years ago, before multi-layer PC boards worked well, there was something called "semi-automated wire wrap". Production wire wrap involves wiring up big circuit boards with thousands of wires. Fully automated wire wrap machines were huge and expensive, and manual wire wrap tended to have too many errors. So "semi-automated wire wrap" was developed. Lights indicated the row and column where the wire was to be attached. The position of the hand-held wire wrap gun was monitored through a mechanical linkage, and if it was in the wrong place, pulling the trigger did nothing. Thus, when a wire was attached, it had to be in the right place.
The equipment for this was far simpler than the fully automated machine, so, using low-wage workers, it became a common way of building boards. It totally de-skills the job. In an hour, anyone can learn it.
It makes me wonder, when this technology is going to enter the field of medicine? I don't think I am ready for a physician with a Dell box strapped to his head..
Harpo Tunnel Syndrome--my wrist feels funny.
The real application for these is an on-demand assembly line to replace a multitude of expensive single product lines that probably don't need to run all the time.
Picture an automotive assembly line that has 300 assembly stations, each one of which gets the "next part" supplied by a chain driven conveyer bringing it to the station on a hook.
Jane, who takes care of placing and tightening down the intake manifold on the engine block in front of her, no longer has to either think about what torque to use, what bolt pattern, or really, anything. Follow the instructions, tighten the bolts per the visual overlay pattern at the designated torque, and on to the next block coming down the line and intake manifold coming off the parts conveyer belt.
The ultimate end of this is much like the Microsoft commercial where the guy in the showroom is picking whether he wants a black car or a red car, and the manufacturing plant is responding almost instantly. Now extend this to not beginning the production on a car until an order is placed, and it'll be ready that day for delivery to the customer's city.
And yes, this reduces Jane to a non-thinking bio-machine for the assembly line. That's the really awful part of this process.
Cast in the Name of Efficiency, Ye Not Cognizant.
Big Ugh.
In the car plant I work at (which will have to remain nameless) the workers learn how to build cars by trial and error.
When a new model comes they start by producing only one a day. The cars that result from the first months of production are so bad that they have to be repaired by experts in a special hall, sometimes taking several days for each car.
The first hundred or so cars are only used for presentations, road tests and crash tests anyway, so it's no big deal if they don't look perfect.
Unfortunately, by the time the car goes into full production most of the workers still don't know what they are doing, and it takes a few hundred defect cars in a row before anyone decide to do anything about the problem.
I guess a system like this would be ideal for the starting phase of production, to train up the workers. The only problem is that whoever sets up the system in the beginning would have to know how to build the car in the optimal way (including all variations). Usually nobody has this knowledge until after the fact.
The other method is active focus which bounces an infrared beam off the object and measures the distance.
Neither is perfect, but in a controlled manufacturing setting it would be easy to create 'focus points' on objects which would allow the system to focus the right distance. Heck, with that you could probably build the system to focus specifically on the correct part, further eliminating confusion.
-