So far there are 2 biofeedback algorithms in the this contraption although this is by no means all that there could be.
One is a straight feedback system that simply adds the input form the sensor to a pre-set in a way which will have no overflow control. This means that with zero pre-set the output will be positively proportional to input. With a little pre-set, less force is needed to reach high power and the motor can be kept ticking over. With a good bit of pre-set then to get max output muscle action must be maintained below a threshold or there will be an overflow and the thing will stop until the muscle relaxes again. finally with a lot of pre-set then the device will be at max at rest and the more force excreted the vibrations will drop off and then ramp up with pressure.
The second will have a time delay in it so it will ramp up slowly, then once a muscle pulse larger than a pre-set value comes along it will drop off, wait and then start ramping again. This can be tweaked in a number of ways e.g. the dead time can be modified based upon the frequency of pulses received or the pre-set could vary with time, either increasing or decreasing.
Of course data can be fed back and forth to the computer which will be able to carry out more complex algorithms which opens up amazing possibilities….
20/09/2008 at 09:08 pm
Peter,
Congrats on getting your system working. One aspect I’m interested in for such toys is using “limited randomness” as part of such a toy. By limited randomness, I mean that the response isn’t totally random, but neither is it so constant (or such a clear pattern) that a user can get entirely used to it. So (for example) the duty cycle might be random, but only within a range that’s chosen — for that stage in use — pleasurable and/or maddeningly teasing.
Another possible use of randomness is to have your CPU “roll the dice” at the beginning, and decide which (of several pre-programmed) scenarios it will run *without* telling the subject which one it has selected. So the subject doesn’t know whether they’re getting a quicky or a long-slow tantalization.
A last note: if you’re not happy with your pressure sensor, here’s a company that makes amazingly sensitive (and pre-amplified) ones: http://www.allsensors.com/
– for example, http://www.allsensors.com/datasheets/commercial_temp/DS-0032_RevA.pdf
measures fractional inches H2O — thus even smaller fraction of a PSI. (Sorry for the barbarian units.)
HTH,
Larry
20/09/2008 at 10:34 pm
I was going to have a go at a low pass filtered random system where the random number only dictates how big and in which direction the next jump will be. If the max size of a jump is significantly smaller than the range of outputs then the device will effectively wander up and down the range.
i like the “lucky dip” idea.
There are all sorts of other possibilities but this is why i have a bootloader so that new patterns and functions can be added at a later date.