Minister of Silly Walks
People Skills: 3517
Registration date: 2009-03-14
Location: Miami, FL / Atlanta, GA
|Subject: Brain-Computer Interface Controls Sun Apr 05, 2009 4:44 pm|| |
- Jeff L wrote:
- Brain-Computer Interface Controls
Despite the exotic name, commercial Brain-Computer Interfaces (hereafter called BCIs) already exist - and you don't even need to shove wires into your head to use them! One model made by OCZ called the Neural Impulse Activator, or NIA for short, is used in conjunction with a standard PC's interface devices (mouse & keyboard). That particular model can be set so certain electrical impulses it picks up are bound to any keystroke the user desires, and can [with practice] result in dramatic drops in reaction time. The number of commands you use the NIA for is also quite flexible, and can it can even just be used as the trigger in FPS games with all other controls kept standard. For more information on OCZ's NIA, go here:
I think these interfaces could potentially be of great use in 'bots. If anything else, the shorter reaction times alone would be reason enough to look into them. However, making a BCI device work with our current radio transmitters would take a great deal of time and effort in programming, troubleshooting, and modifications to a transmitter's wiring to allow it to be fed information by a computer hooked up to a BCI.
In any case, this is at least a very interesting idea that will probably become reality in a few years. In fact, several toys are already on the market that use simpler BCIs (mostly Star Wars branded).
This is considerably more challenging than your typical upgrade, but is not outside the realm of possibility. I've wanted to make a robotics platform that works with these headsets since I first heard about them back in '07, and I may soon have the resources to be able to start in on work for a simple two-channel (possibly three or four channel) interface.
Challenges to Overcome:
1. Block EM Interference - the headsets use electroencephalography to detect a variety of electrical nerve impulses on the surface of your forehead or scalp. Since they use very picky sensors, electrical noise from anything ranging from a cellphone to a charging brick can mess with them. The first challenge to overcome is to find a way to shield them from any external interference to improve reliability.
2. Make a USB Adapter for Spektrum Transmitters - This is going to be one of the more difficult steps in developing the platform. This boils down to making some kind of bypass around the potentiometers and switches in the transmitter to allow a computer to determine the values being transmitted to the robot instead.
3. Make a Converter Program that Works with the Neural Headset's Drivers to Send Values Over to the Transmitter - This is surprisingly probably going to be one of the easier steps. Since the headsets all interpret voltages as keystrokes or mouse clicks, you just have to make a program that sends the right information to the transmitter with each virtual keystroke or click.
Estimated Development Cost: $200 for the headset, $300-400 for the computer, $200 for the radio, $50 in miscellaneous electrical components - $850 in materials, Total. Time Required: Around three to four weeks of solid work on just this project alone, minimum. Could more reasonably have a rough prototype with several months worth of work.
-Make the converter program be able to send a range of values, not just all ON or OFF
-Wide-angle camera system (maybe fish eye lens?) built into the robot to give the driver the option of a more immersive environment [while we're at it, if the camera was made you could potentially make a HUD at a later time with sensors transmitting back information on battery status, motor RPMs & temps, ESC temps, etc. But that's getting a bit ahead of ourselves, and may require too much weight even for a 120 pound robot]