I am the lead technical engineer at Nuvation on this project. I just wanted to clear up a few things we have seen on some websites.

It is correct that Nuvation didn’t build the industrial robot arm itself; however we did build a custom processor interface board which connects together the machine vision system, Coldfire processor, and Robot Arm. Nuvation is a full-service electronics and software design-house, and we are capable of a wide variety of technical services, including complex PCB design (analog/digital/mixed-signal), FPGA/logic design, embedded software design, RF and signal integrity, and turn-key product design — all the way through manufacturing. We also help a lot of companies architect new products, or troubleshoot existing products. The Air Hockey Robot was a special project for a tradeshow.

And yes, I can confirm that we completed this project in just over 2 months, mainly because we were asked to do this project about 2.5 months before the tradeshow. I will admit there were several long work weeks however. :)

If you have any other questions on the demo or Nuvation, feel free to email me.

mohan.gurunathan@nuvation.com

As far as the ‘AI’ goes, I’d say they need to look into more organic programming, set the basic rules (i.e. don’t let puck enter own goal, reject puck on trajectory to other goal…etc) and let it find the best way of doing things statistically :)

Ahem… Would’ve taken them longer than 2 months though and as it’s only a POC / showcase probably not cost-effective.

Good stuff anyway :)

not sure why, but i believe this worthy.

One is detecting white disc on blue background using a camera located directly above. I don’t thing it’s anything very sophisticated for today’s standards.

Next we have a game algorithm. Again, making a computer play air hockey shouldn’t be a rocket science. On the video you can see that computer actually sucks at putting the disk into opponent’s net (come one, you can do better than that).

And finally we have the robot (arm).

Not very spectacular but still nice stuff. Nice as a project “done because I can” or to gain some experience but as a research I’d say it’s nothing spectacular.