IBM programmer controls a BB-8 droid from 'Star Wars' with his mind

This is cool! The video made me chuckle because it's nearly content free but has a lot of techy jargon, and because of the shots of him outstretching his arm using "the force".

The tech is real though, even if they didn't say anything meaningful about how it's actually done.

I saw a research paper talk at the EVOSTAR conference in 2007, and the author had successfully controlled a mouse cursor in 2d using a handful of sensors placed on his head, and training a neural network. He couldn't move the pointer quickly or particularly accurately, but it definitely works. He also mentioned that using this contraption and training the software was the single most mentally demanding thing he'd ever done, and that he'd be completely exhausted and sweating from his scalp after 7-8 minutes of trying.

*edited for spelling

I've tried the Emotiv headset, it doesn't work as advertised at all. Occasionally I could get the device to send the signal I was imagining, but it seemed pretty random. This is pure marketing for IBM cloud services, the video doesn't show or prove that he can move the device with is mind with any degree of control.

Emotiv did more of the work than IBM here. Watch Tan Le's TED Talks for more details.

I couldn't tell from the article or the video what the latency of this thing was. What is the error rate in interpreting his states of mind and translating them into commands to the toy? The demonstration would have been much more impressive if he could show a video of the robot performing complex maneuvers from thought control.

Anybody tinkered with the headset? It's $360 for the most basic.

https://emotiv.com/insight.php

EDIT: seems you need to buy the $750 to read raw data.. and you can only publish open source apps or you need to buy a $1000 license, includes 5 user at 1 platform

If you want a cheaper consumer EEG sensor, you can check out Neurosky, which retails ~$100 and also has an SDK. Be forewarned about consumer EEG headsets though: they're slightly less random than a PRNG. Real EEG experiments usually happen in Faraday cages.