Enhanced energy capacity from a small battery is an interesting achievement, but I am more interested in how the mind controls a robot or anything else. Can you elucidate for me how this is accomplished? At least one enquiring mind would like to know.
Ratch
No problem Ratch, it's no secret.
There are a number of different types of simple EEG headset available. They all use a "TGAM1" chip, from Neurosky. The book uses a Neurosky Mindwave headset. They're a bit expensive for me, so I opted to use a headset from the Mattel Mindflex game. As mentioned, they all use the same chip, but a micro in the headset parses the original data stream and selectively decides what to send.
The original data is Poor Signal, ( inverse of signal strength, where 200=headset not connected), Attention, Meditation, Delta, Theta, Low Alpha, High Alpha, Low Beta, High Beta, Low Gamma and Mid Gamma.
Poor Signal, Attention and Meditation are 8-bit, and the others are all 32-bit values.
In the case of the Mindflex, (which I know the most about), only the Poor Signal and Attention values are actually sent to the base board by the headset micro, using a 2.4GHz RF transmitter.
That game uses the Attention value to control the speed of a small DC fan on the baseboard, which blows a foam ball up into the air. Also on the baseboard, there's a small, manually-operated turntable that's used to rotate the fan position so that the user can manipulate the foam ball through an obstacle course with height controlled by their brain waves.
The other data is still available at the TGAM1 chip in the headset, though, so I'm hacking the chip itself, then using my own 433MHz RF module to transmit the full range of data to an Arduino, where it's parsed to extract all of the values, for (hopefully) more advanced control.
By default, the TGAM1 chip sends data out once per second at 9600 baud, but it can be reconfigured to send all of the above data once per second plus the full raw data every 7mS at 57600 baud.
After I get it working properly at the once per second rate, I plan to reconfigure the chip so that I have access to the full raw data at the faster rate. (Once per second is a bit slow for decent control applications, I reckon.) The chip can be reconfigured by a software command, which resets to default when the power is cycled, or by hard-wiring on the headset board, which is then permanent.
There's a fair bit of info available on the subject on the web, but if you have further questions, fire away.....
I should add, the book's method also only gets the Attention value, to use for controlling the speed of a pair of continuous rotation servos.
Edit:
@Harald Kapp , perhaps this thread shouldn't be moved to the comedy thread after all. It's developing into something more than a joke.
