Nice idea... quite low tech, and it needs a WBTB, but sounds great.
I'd play with a vibration on my wrist instead, but same thing applies.
Hmm. Anyone gonna try this besides me?
Printable View
Nice idea... quite low tech, and it needs a WBTB, but sounds great.
I'd play with a vibration on my wrist instead, but same thing applies.
Hmm. Anyone gonna try this besides me?
Well, since you have a vibrator on your wrist, you could place one contact there as well, a small metal object could be the other contact, forming your switch.
When you detect a drastic change in either resistance or capacitance between the two contact, then you know the person has fallen asleep. Let me search for a few simple circuits. This is much easier than the sensitive electro-galvanic stuff we were doing earlier.
http://www.discovercircuits.com/S/s-capacitance.htm
Hmm, maybe easier would be to hold a mercury switch in you hand. When you relax, the hand falls over tripping the switch.
So many options, so little time....
Uh just to add to the mix:
- It has to count down after the trigger is released - eg. 5 mins
- It should preferably only vibrate for 1 second, then stop.
- And preferably repeat that 1 sec eg 3 times, after eg. 5 secs
That makes it a tiny bit trickier, but as you say a whole lot easier than before.
You could build this into a gardening glove, with a small battery and cellphone vibrator (I have a broken cellphone)
[EDIT: The main reason I like this idea btw, is because it should be quite easy. I still would obviously like to figure out how to detect REM with electrodermal response or blood at a later stage]
So easy and could be self contained if built in a garden glove. Sure beats being wired to the computer. Or course, the switch could be made optional. Just wake up and arm the thing. i usually fall back asleep a few seconds after waking at 6am. This is my WILD time anyway and I;ve had a horrible time staying awake through it lately.
Could somebody perhaps write a short summary of the ideas in this thread? There's a lot of good ideas here but reading through all 9 pages is a bit much.
Anyway I really like the dead-man switch idea. I might try it now that there's a decent electronics shop nearby.
Yes the switch is an interesting idea, however there are many stages of sleep, not all of them are REM, so the flashing lights 5 min after you fell asleap would be useless in many of the stages of sleep. Of course its a great idea if you are in the right stage of sleep and you have it timed right.
[example of sleep stages]
http://www.sleepdisorderchannel.net/stages/
muscles dont always relax after you fall asleep. i have seen my sister dead asleep on the couch, holding a huge biology textbook upright in her hand, which was over the edge of the couch. it was funny looking :lol:
i think the best way to detect rem is sensitive heart monitors, or actual eye movement, the latter being the most reliable.
I do have the idea of using an optical mouse to detect your moving eyes... I am working on a circuit that takes out the annoying red LED when it isnt needed, you could even use a wireless mouse to send eye movement data to the computer...
Exactly, the idea was to go to sleep for 6 hours first, so that you are far more likely to be in REM within a few minutes of going back to sleep.Quote:
Originally posted by Uncharted
Yes the switch is an interesting idea, however there are many stages of sleep, not all of them are REM, so the flashing lights 5 min after you fell asleap would be useless in many of the stages of sleep. *Of course its a great idea if you are in the right stage of sleep and you have it timed right.
*[example of sleep stages]
* * * http://www.sleepdisorderchannel.net/stages/
You wouldn't be able to use this idea to go to sleep in the evening - the signal would trigger in delta phase (deep sleep)
Thanks kimpossible ;)
Yes, I'm very interested in your progress. You seem to have a lot more resources available to persue this, and I'm eager to see how it goes.
As for me, realistically I can only get into things properly again from about Jan 2006
I really need to get things settled down on my end :)
One big advantage you have is that you are far more knowledgeable on electronics than I.
So you're going to try a EEG? Should be interesting, but I worry that it will be quite intrusive.
Please keep us up to date ;)
Sounds exactly as I had in mind originally, before the dead man's hand compromise
The difference is that you have the resources and knowledge I do not ;)
Once again, PLEASE keep me up to date. After all the original drive was begun by me, so I'm intensely interested in the outcome and progress - even if I'm not in an immediate position to do it myself.
This circuit looks to me like it will deliver an analog voltage to the PIC. What Placebo and I have been looking at originally was either a PWM or a VCO that delivered square waves to the CTS pin on the RS232 port. Do you have a cheap to build circuit that does this?
Perhaps we can use the circuit you provided as the frontend to the VCO.
are you useing tubes? aren't those hard to find nowadays?
cd
Gotcha! Man, I've got a whole stack of micro industrial controllers sitting in front of me that would be perfect for this. Unfortunately, retail on the cheapest of them is around $200.Quote:
Originally posted by kimpossible
Yup, it's delivering analog (ma) to the pic. *The pic I'm using (picaxe 18X) is a DIRT CHEAP part with 3x12bit a/d *8xD (I/O) and seriel out.
BTW, you never answered my PM.
haven't answered mine either.....batteries huh? i didn't think they put circles around em liek that. oh well, they jsut seemed to resemble tubes, like cathode stuff.
cd
Hello all! I went through most messages and got an overview of your project. At the moment I'm not too excited about the method of your choice, because for an easy project it sure seems to take a lot of effort and I don't think it's necessary that reliable. I want to pursue PIR and tactical/visual (possibly audio too) cueing.
Does anyone have sources for suitable PIR sensors and Fresnel lenses? I'll try to contact someone who has a Remdreamer and ask them what parts they used for starters.
Don't let me discourage you, stay on it, because if it will work out, the visorless design will be pretty cool.
Weird... well, good news then. I have a GSR machine with computer output, I suppose I could somehow use it then? You said you had some software ready, maybe I could try it. Do you have the source online?Quote:
Originally posted by kimpossible
It does work, and rather well. *Atleast for me. *My neighbor started testing it last night, so I should know more there shortly...
So you're saying that it's impossible to hold the mask in place.. That novadreamer anecdote sounds weird. How don't people return their units if they work that badly? (are they really happy to pay $600 for something that doesn't work). The reason could be that their PIR was useless.
What kind of vibration you use in the glove?
Did you get any results?
I'm electrician and programmer , but newbie in psihology and lucid dreaming.
I'm planing to make a device with microcontroller that will log skin resistance at night.
I will post results and hope you'll help me to understand them.
Hello Nunika!Quote:
Originally posted by Nunika
Did you get any results? *
I'm electrician and programmer , but newbie in psihology and lucid dreaming.
I'm planing to make a device with microcontroller that will log skin resistance at night. *
I will post results and hope you'll help me to understand them.
I see this is your first post! Welcome to Deam Views And the research Team!
To give your efforts due credit you should begin your own post in regaurds to your research.
This Experiment may help you a lot. There is a ton of information in here.
I removed most of the specifics of my efforts as I'm planning to build the device commercially.
Thats good ! I think. It seems to work at least.Best wishes! And dont get greedy!Quote:
Originally posted by kimpossible
I removed most of the specifics of my efforts as I'm planning to build the device commercially.
:wink:
I GOT IT ! :) ... well, he did ;)
all we needed was to wait... until SOMEBODY ELSE builds it. :D
http://www.fuzzymonkey.org/newfuzzy/softwa...inheartmonitor/
after re-reading the thread i realized how uselessly complicated the design i was going for was and i started reading about the 555 chip... THEN i realized heart rate... though i don't think it changes significantly with REM sleep, might be a good experiment. so i started searching on google, and eventually found the link above. it provides BOTH the interface AND the software so, if heart rate gives no results in terms or REM detection, we can slowly toy with both the hardware and the software until we get a reading of electrodermal activity (which i know for sure changes drastically... or at least i think i know for sure ;) hehe).
point being:
1- this is simple to build
2- i can test it right away after it's built
3- it basically measures electrodermal activity at two points, eliminates the noise (from point 3?) and outputs the clean pulse voltage.... if we compare the voltage from the two probes with a CONSTANT one, we have electrodermal activity (i think?)... someone please let me know if i'm wrong... i think this is right though... the "noise" varies throughout the day, and this is what we actually want to monitor, the average noise and not the heart rate...
4- there is already software built for it, we can just modify it or add to the features (which i don't think the developper would mind at all)
if it doesn't work, we can toy with the software so that we get a more general reading... and then we can butcher the hardware and monitor sleep... until we notice some significant changes... or someone with an understanding of medical science and electronics sheds some light on what exactly it is that we are doing :)
i'll get back on this (i have some, though not a lot, spare time to toy with... i'll try to buy the parts i need this friday and get going with it)
My experiements with heart-rate (5 people) to detect REM were abject failures. Sensitive enough and you couldn't eliminate the noise... You're looking at sub 0.1%. Maybe really good FFT software on a multi-processor machine. . .
yeah, heartbeat is really crap at detecting rem... it goes down a bit during regular sleep, and it's comparable to waking states during rem, but the difference is very small... not to mention the fact heartrate cycles render any kind of monitoring (other than average over periods of time) completely useless...
i just realized that now. AND i also realized that the diagram uses an amplifier that looks at very small differences and amplifies them... if i were to take out one of the values, i'd have a HUGE difference, so the amplifier becomes somewhat pointless, i could probably do with a more generic amplifier.
i still think i'll buy that gsr2 and see how it works, but i'm afraid it won't be very useful either... i'm starting to have doubts when it comes to these type of measurements for a few reasons: they require good electrodes... i tryed toying with home-made ones and they just don't do the trick... they're not very reliable. not only that, but the electrodes have to be fairly fixed, meaning you're looking at tape or whatever way of fixing them, and slight changes would offset the calibration... to add to that, the entire process of putting them on/taking them off seems somewhat annoying, especially if you plan to do it fairly often.
so i dunno...
i'm starting to think the best and surest way is still IR detection... it's called RAPID EYE movement because that's the most noticable feature of the sleep period...
i'm curious as to how your device works kimpossible (not specifically, if you're planning to release it comercially, but as an overview) and how that proved to be effective (are there no problems with different values from person to person, etc?)
My GSR system has had amazing accuracy across six people. Compared to EEG and Camera (two highly accurate systems), I run 97%+ accurate detection over 300+ nights of sleep, with zero false-positive detection. (I miss some, but I don't detect what isn't.)
I run some pretty heavy-duty processing on the built-in processor - including FFTs to get that kinda accuracy, though
well, with that accuracy... it sounds awesome.
...
when is it being released ? :D