Torso position determined via accelerometer!

avi123 wrote:Max, please don't tell me that you compete with SAS and Jim Goodnight?

We frequently (um, I think always) out perform SAS in head-to-head customer evaluations, both in time to, and quality of, the solution, usually in the marketing sector, but we rarely compete because we don't come from a statistical perspective (theoretical), but an engineering one (practical). You wouldn't run/optimize an oil platform with SAS, for example, maybe with Aspen. Our technologies are automated and self-optimizing. Theirs often have to be "coded", which takes time and (even more) money. BTW, going back to "where the money is", you do realize they are private as well.

Still, I read here that Jim Goodman is worth $7.2 billions:

http://www.celebritynetworth.com/riches ... net-worth/

Is it not where the money is?

avi123 wrote:Is it not where the money is?

EXACTLY my point earlier! They are a private corporation. We are a private corporation. Earlier you said, "Max, from your post I can tell that you are not a public company where the money is." My point is you can be "where the money is" and be PRIVATE. Thank you.

Done !

Nice unit, BasementDwellingGeek! My X16 1C accelerometer came a few hours ago. We created a data importer (merely modified a copy of an existing one) to gather any number of files from a folder, parse them, delete or archive them, time order the blocks of data in case the files happen to be out of order time-wise 'cuz you can't go by file name you mentioned, fully and properly time stamp the data and import the XYZ series into our historian (or pass the data on to other processes).

I did a "put it on the desk and flip-flop it" and sucked in the data and then extracted it again and demonstrate a chart:



Now all I have to do is in the evening, push a button on the device and wear it. In the morning I merely drag-drop the files to a shared folder... Slurrrp... in the data goes, just like I do with my ResMed and Pulse-Ox data. Easy as pie. I'm debating a waistband thingy or maybe attaching it to my mask strap. I may decide to post-process the data into positions, easy enough to do, but I probably can deduce what's going on just by looking at the XYZ data.

Now I have my ResMed, Zeo brain waves, pulse-ox, and accelerometer data all together, all properly time stamped, and fully able to analyze it. This should show me arousals too (NO, not THAT kind... sheesh). WOOT ! THANK YOU BDG!

From previous posts I figured you wouldn't have much of a problem. Have you experimented with the real time aspect? I have noticed clock creep. Supposedly there is a driver available to reset the clock or just use the supplied application.

I have experimented with a few different mounting options. Depending on how the device is oriented in its mounting and where the normally up position is desired to be, different axes (plural of axis, I looked it up) are required to calculate the angle needed to properly display the orientation symbol.

The first mounting configuration I tried was strapped to my headgear. There is a band that goes over the top of my head. I tied on with the USB port to the left and the start button to the right. I later used that same configuration with the device taped to my abdomen just below my ribs. In that configuration the X axis became the primary with the Z axis the secondary. Taping every night and ripping the tape off only to throw it away the next morning was seen as a nuisance and unnecessary recurring expense.

Fast forward to mounting experiment #3. I stopped at the local Walgreens and picked up a 4 inch self adhering "Ace Bandage". Pretty weird stuff with a slightly rubbery feel, and no clips. The package states "No Latex". No length was given but it measures about 56 inches in a relaxed state. I was able to wrapped it 3 times around my chest and tucked the tail end under the initial at the end. I inserted the device between the first and second wraps in a vertical orientation, USP port up and flat side against my chest, hoping it would stay in place all night. It did. In this orientation the primary axis became the Y axis with Z as the secondary axis.

A note about primary and secondary: A single axis can give tilt but not whether one is face up or face down. The secondary axis is required to make that determination.

In anticipation of overnight experiment #3 last night I spent a few hours with my stripped down "finger this out" project yesterday evening and updated the complete application this morning. I added controls to select the primary and secondary axes, formerly AX and AZ. Now I can choose any axis as the primary and any OTHER axis as the secondary. There are additional controls to "fine tune" the algorithm.

Yes, mounting is my next challenge. I merely stuck it in my waistband and that's not stable enough to be consistent in orientation, but I did get the positional transitions, the movement events. I don't think I got the specific orientations because the device shifted its orientation during the night. The Walgreen's vet-wrap bandage is an idea. Another idea I thought of was making a velcro pouch / pocket, hook side out, fuzzy side in, that I could stick to the mask strap that goes across the top of my head since those straps are fuzzy and velcro friendly. That way I could open an end of the pocket and slip in/out the device.

Here's what I got last night, tho the device probably rotated about some...

OK, this helps understand what is going on... Tho I do not yet have true position, I can see a story in the changes in XYZ vs events and mask pressures.

Top chart is accelerometer crude positional changes, then my ResMed breathing and mask pressure, at the bottom is the Zeo Sleep Stages.

(1) I was coming out of deep heading into REM and started having some events, so I repositioned a bit and that helped.
(2) I was near the end of REM, shifted a little a few times, produced some events, must have found an OK, not too different position and events subsided.
(3) I was shifting from deep to light sleep, repositioned, but was not helpful and events fired up and mask pressure rose to compensate.
(4) I entered REM and fired a bunch of events (probably on my back), I repositioned, resolved it and mask pressure fell off accordingly.
(5) I repositioned and fired off a false central, just like people say we do.
After 5, the shifting positions was me awakening and looking at the clock, deciding to keep sleeping or get up.

I think this will enable me to count the times I am sleep disturbed/shift/arouse. The count per night probably is related to my "sleep quality" and maybe next day feel.

A crude assessment, first go at it. I'll make some technical / mechanical adjustments.

MaxDarkside wrote:Yes, mounting is my next challenge ... Another idea I thought of was making a velcro pouch / pocket, hook side out, fuzzy side in, that I could stick to the mask strap that goes across the top of my head...

I found a small piece of velcro from a S&W 9mm nylon pistol holster I'm not using at the moment ... ... and temporarily rubber-banded the Velcro to the accelerometer and attached it to my Q-FX mask strap on the top. Seems snug enough for a trial. If it does not fall off during the night, I'll make a permanent pocket-like Velcro holder, leaving a spot to "push the button". If it works, I don't even need to detach it from the head gear, just plug my head gear into the computer in the morning when I wake up !!

MacGyver would be proud. LOL

Edit: I realize that my head could be left or right and still be on my tummy. Hopefully if I'm face up, I'm on my back

MaxDarkside wrote:OK, this helps understand what is going on... Tho I do not yet have true position, I can see a story in the changes in XYZ vs events and mask pressures.



think this will enable me to count the times I am sleep disturbed/shift/arouse. The count per night probably is related to my "sleep quality" and maybe next day feel.

Question,

Assuming that you find out about it, how would you decide that this and not something else caused your sleep disturbance that night?

avi123 wrote:Assuming that you find out about it, how would you decide that this and not something else caused your sleep disturbance that night?

This measures movement, an effect, not a cause. It may help me quantify disturbances. Finding out the cause of those disturbances is another matter.

The velcro-ing of the accelerometer to the top mask strap worked pretty well. As a calibration test, so I could interpret XYZ in the AM, I laid on my tummy head to left for a bit, then rolled on my back face up for a bit so I can see the related XYZ values, then rolled on my tummy again and went to sleep. What I saw in the morning was pretty clear. As soon as I rolled on my back mid-sleep a cluster of events erupted until the machine could raise pressure and eliminate them. As the machine started dropping pressure after the cluster was tackled, when the pressure dropped to a certain level a few events started showing again. Just at that time I rolled onto my tummy again and I had clean sleep until I rolled on my back again some hours later and erupted in events again. The machine started raising pressure, I shortly rolled on my tummy and the events subsided. etc, etc...

I know I have positional apnea. Nothing new there, but this makes it obvious. I can also see what minimum pressure is required to avoid apneas when on my back (new knowledge)... about 10 or 11 cm H2O, but need to go to about 15 to eliminate clusters once they start.

I also learned it's merely being on my back that causes my apneas. I thought it was on my back while in REM.

This is useful. Again, thanks BasementDwellingGeek.

I love it when a plan comes together! I think it was you that tuned me onto Zeo. I'm glad I could reciprocate.

I was surprised at how much time I spend on my back, since I am usually on my hip or belly when waking. Expiratory mouth breathing is worse while on my back. Time to start thinking about chin strap.

Yes, BDG, I would say for me this accelerometer information is as important as the data from my Zeo, maybe more so. Below is a shot of my events, mask pressure and my positions, with "on my back" noted by the high red line on the accel graph at the bottom, inside the blue boxes. Being on my back explains 100% of my apnea events. Now to seek out a large Pugsy-recommended buck-wheat pillow to see if I can prevent or reduce my rolling over...



I'll come up with a scheme to chompulate the XYZ transitions and sum their quantity per night as a metric of "restfulness" or "sleep tranquility". Some movement is necessary for fluid dynamics in the body, but last night the number of my position changes were higher than the others.

MaxDarkside wrote:Now to seek out a large Pugsy-recommended buck-wheat pillow to see if I can prevent or reduce my rolling over...

I am leaning towards the small body pillow offered at 72 beans. I have found that the Regenisis pillow (not a large one though) will move out of position sometimes...though it stays in place better than a conventional bed pillow.
A king size buckwheat pillow would work great as a back brace as long as no one else is in the bed.
I have not tried the Large buckwheat Regenisis pillow to see how much of a bed hog it might be...for sure a long skinny buckwheat pillow would be less of a bed hog. Hence my reasoning for leaning towards the small body pillow. I feel the large body pillow would be too much of a bed hog.

I have not ordered the small body pillow yet but I suspect I will at some time in the future. I just spent my mad money on new copper rain chains for down spouts for my new gutters...so have to wait just a bit. I could buy 4 body pillows for what I just spent so best wait just a bit. I don't care though. I love what I did..and I am worth it...now if we could just get some rain. All around us last night and barely a sprinkle here.

Adding more buckwheat hulls to the Regenisis pillow I have might just work. I had to remove a substantial amount from the king sized pillow I got from 72beans...I need just a little more weight to keep the Regenisis from sliding around part of the time (sometimes it doesn't move but sometimes it does)...so I will add those hulls to the Regenisis to see if it works.
FWIW I just measured the Regenisis pillow 18" X 14" X 2" in case anyone was wondering. It is not very large. Remember, this is the original size that I received long ago as part of a free trial before Bret improved size and weight. It suits my head and neck because I am not very big either.

All you really need is a dog that loves to snuggle! When I turn onto my side, my dog slams his little body lengthwise along my back and I can't move without squashing him so, needless to say, I don't. It's funny when something I used to find kind of irritating turns into a blessing in disguise.

Suz