yay accelerometers

[gfish]

As anyone who has slept with me will confirm, I tend to toss and turn all night. Thanks to a fun Android app, even the distressing number of you who have little desire to share a bed with me can see just how bad it is:



I only used it for the first time last night, but I'm curious to see what patterns emerge over time. I'm looking forward to playing with the optimally timed alarm feature -- I've wanted to try that for years, but custom hardware is expensive. Yet another reason that having a general-purpose computing platform with an accelerometer is awesome!

Comments

[hsifyppah.livejournal.com]

Man the sleep debt tracking option looks super depressing.

I've seen complaints about a best-selling iphone equivalent that if you leave it motionless on a table, it will also generate an obviously bogus sleep cycle graph and awaken you at a random time that it claims is in light sleep. Something to test with this app before you get too involved in data collection!

I was delighted to find out the other day that Cedar Point (big amusement park) gets enough high school physics day type groups that they have a really strict accelerometer policy. Hee hee hee. They require wrist straps, only one style is allowed, and they're banned on certain rides. I'm picturing dodgy science rebels sneaking accelerometers on the banned rides. Or maybe making a big show of leaving their accelerometer in the bag check bin, and pulling out a huge old-school barometer in a mahogany case and saying "This is cool, though, right?"

[gfish.livejournal.com]

I left it on the bedside table last night, and the graph is completely flat except for a handful of spikes which could easily have been me bumping the rather wobbly table. I'll leave it on the windowsill to confirm this, but it looks pretty good.

[tereshkova2001.livejournal.com]

A friend of mine has an iPhone version and really likes the alarm feature. I'm not sure how useful it is if you have a sleepIng companion, though.

[cow.livejournal.com]

Hm. I am reading up on how it works, because it seems neat. Mostly, I'd be afraid of smashing my phone or knocking it off onto the floor, because I also toss and turn quite a bit.

That said, it's hard to imagine you going to sleep at 23:09. :3

[gfish.livejournal.com]

I know! And that was kind of late. I don't even know who I am anymore! :(

[randomdreams.livejournal.com]

That's really cool! What a great idea. I've read about scads of neat distributed data-gathering projects using those accelerometers, but this one is thoroughly sweet. (I'm working on making a vibration spectrum analyzer right now so I've been playing with some, but none as of yet that are as sensitive as the one in the android -- albeit they can sample a lot faster.)
My brother moves extensively when he sleeps: he's cracked the welds on bedframes, regularly awakens on the floor rather than in the bed, and has to leave the bed well clear of the wall for fear of trashing the drywall. In marked contrast, I sleep in exactly the same position, without moving, for roughly 3 hours, wake up, turn over, and go back to sleep. However, with that said, I'd love to see what the android analyzer says, and see if I'm actually as still as I think/have been told.

[ljtourist.livejournal.com]

I need to get this immediately if not sooner! Thank you for bringing it to my attention.

[zzyzx-xyzzy.livejournal.com]

Hmm. it said it was going to vibrate but didn't. Relatedly, is there any way on android to use the builtin speaker even when a headset is plugged in?

[gfish.livejournal.com]

There might be an app which allows that. I have one which will disable the headset control buttons, for instance, so when I plug in non-standard headsets it doesn't get random pause and skip signals.

[gustavolacerda.livejournal.com]

Thanks for the tip. I got the app and tried it. It's really really well done!

I think there is a fundamental problem with setting the threshold... the phone doesn't know how close it is to your body. In my case, it only detected actual awakenings (two big spikes in the morning), with much smaller bumps at ~6:15am and ~3:45am. So it's unclear from your graph how many times the threshold was crossed (and how reasonable that threshold is).