World's Fastest Robot Versus the Wiimote

kkleiner writes "Adept's Quattro, a placement and sorting arm, took the title of fastest robot last year, but it was only recently during National Robotics Week that it met its most gruesome opponents: nerds with Wiimotes. Visitors tried to keep the Quattro from placing and sorting on a small mechanized platform by moving it using the Nintendo video controller. The bottom line is that when it comes to simplified and repetitive tasks there's really no beating robotic prowess."

Comments

They're pretty good at working on humans, too

[Bearhouse]

See:

http://en.wikipedia.org/wiki/Robotic_surgery

When my mother had her hip replaced, the surgeon said that robot-replaced hips lasted longer due to better hole size and placement, (they make a hole in your bones and hammer the replacement joint in...)

As for this kinda flaky 'robots vs. humans' story:
1. We'll never be able to beat a robot's reaction times {see note} speed and/or raw power, but
2. Until AI improves, we'll still be the ones programming the things

Note: What was that SciFi story about humans being 'paired' with cats in order to have both high intelligence and inhumanly fast reaction times?

really impressive

[pz]

Go to the linked article (yes, yes, do it anyway). Skip the Wii demo video that forms the basis of the post because it really isn't interesting. Go to the second video. Watch it.

Holy frick. Robotic vision and control has come a long way.

Re:really impressive

[MozeeToby]

Yeah, but am I the only one here who thinks that a lot of those tasks could have been done with machinery orders of magnitude simpler and cheaper than this robot? Seriously, half their examples of 'real world' usage were moving things from one conveyor to the other, with no sorting or filtering required. Some of their examples (like placing the chocolates in the correct locations in the box) were impressive, but it just felt to me like they were showing off when much simpler designs could have been used.

It has been this way for a while.

[Whatsisname]

It's been known for a long time that robots and computerized systems are vastly superior to humans at simple tasks, their only downside is the upfront cost and often inflexibility.

One of the neatest applications I saw recently was in a factory where macadamia nuts were shelled. The nuts would pass through a big set of rollers, cracking the shells open. Then, the shell casings and the nuts would fall down, and a computer vision system would detect the nuts and the shells. Everything then fell through a collection of compressed air blowers, that would precisely blow the macademia nuts out of the stream of falling shells onto a conveyor platform, while the shells would fall seperated into a hopper off to somewhere else.

Re:So wait

[tlhIngan]

You took a robot, capable of crunching numbers at speeds in excess of a thousand calculations per second, programmed it and engineered it to perform a specific task, and then wanted to see if humans, who take 1/5th of a second just to react, can't do any more than a few SIMPLE calculations in a second, and had them use the worlds laggiest controller, and wanted to see who would win?

It's a pick-and-place machine. Most PnP require that the inputs and outputs are stored in well-known locations, and have pretty basic image recognition software (they can tell if a black blob is in the wrong place, for example - if it was loaded wrong). Or to handle the slight misalignment of the source or destination.

In this case, the robot is picking and placing from and to a platform that can move arbitrarily, while it's even doing the picking and placing. That implies it not only knows it has to look for the source and destination, but recognize the platform and perform the task. Even if the thing it's grabbing suddenly decided to move under it while it's doing the picking or placing.

The human might be slower, but they're also a lot more unpredictable, so the robot has it use up its millions of calculations per second to figure out where things are and react when things start moving from under it.

Re:So will manufacturing return?

[OzPeter]

Whenever I see these reports, it makes me wonder about the implications on manufacturing. Someone in the US or Europe can't/won't compete with someone in China working 15 hour days in a sweatshop for 50 cents an hour, and so from the company's standpoint, it makes economic sense to move. But will the rise in robotics cause a return of manufacturing? You will still need some people working in the factory maintaining the robots and what-not, but it may be cheaper to manufacture things closer to their destination rather than manufacturing them in a developing country and shipping them.

You do realize that manufacturing in the US has been automated for a very very long time and it is *still* going down the drain.

But even with automated manufacturing you still need manual labour to work the production lines. Its economically infeasible to produce a robot that has the flexibility and dexterity of a human for general purpose use. You should watch the shows like "how its made" etc. They show lots of automated processes, yet there are always manual steps involved. I remember seeing one show where a workers job was solely to turn socks inside out.

In the end, US companies have already cut resources to the bone in order to stay competitive with overseas manufacturing. What would really help is de-walmart-ing the consumer part of the equation.

Re:So wait

[SerpentMage]

Ah grasshopper you have achieved the first level... But to reach zen you must go to the next level. And after that one day you might even reach the following:

http://www.youtube.com/watch?v=xNG3sgk02Lc

http://www.youtube.com/watch?v=q395-F6hAcg

I want a machine to beat that! My head just spins thinking, ok so where did that cup move to?