Realtime CPAP Metrics Possible?
Hi all!
TLDR:
I'm currently experimenting with the idea of parsing and streaming CPAP machine metrics (AHI, Flow Rate...etc) in real time as a personal project. The possibility of doing so lends itself to a number of interesting, maybe even useful, applications. I have however hit a huge wall. My original method of using a wireless SD card to access data in the CPAP as it was being written to disk turned out to be unreliable. There must be an internal buffer of some kind, because the new session metrics file was not being flush to disk while it was running, in fact I only ever saw it happen once the CPAP machine was stopped. So I thought I would seek the advice of people here. Is there some kind of cable I can buy to fit the proprietary hardware slot that normally a modem would fit into and maybe pull metrics that way? (Btw all my work is open sourced, so nothing proprietary!).
Anything helps, thanks!
My machine: Philips Respironics REMstar Auto A-Flex (System one), pictures included in post.
The longer version
My name is Dan, a Software Engineer that has recently started taking an interest in having more ownership over my sleep apnea, namely at the software/hardware level. I've was diagnosed with sleep apnea probably 12 years ago. A big turning point for me was finding SleepyHead and realizing just how much information and personal sleep data was being withheld from me. I've read over the source code for SleepyHead and learned a ton from it, I even ported a custom data parser specifically for my machine type. However I felt like there was a next step missing. I wanted my sleep data in real time, there is a lot you could do with that information. What if you had custom machine learning algorithms that took in all of you historical sleep data and could identify predictors for when you'll have an apnea event? You could then use the streaming metrics to adjust certain therapy characteristics (or real world stimulus) in real time for better patient outcome. My current goal is a lot smaller then that though, I actually just want to use it for a creative fun artsy project to raise awareness for sleep apnea. If you haven't already read the TLDR above I've basically hit a wall and have almost given up. However I recently read a VICE article (can't link, new member) about SleepyHead and Mark Watkins that has really fired me up again. There has to be a way to retrieve these metrics in real time. It's done during sleep studies and my hope is that there is a cable for my machine that allows direct access to some internal memory or buffer where I can consume these metrics as the machine processes them. If you know anything please let me know!
TLDR:
I'm currently experimenting with the idea of parsing and streaming CPAP machine metrics (AHI, Flow Rate...etc) in real time as a personal project. The possibility of doing so lends itself to a number of interesting, maybe even useful, applications. I have however hit a huge wall. My original method of using a wireless SD card to access data in the CPAP as it was being written to disk turned out to be unreliable. There must be an internal buffer of some kind, because the new session metrics file was not being flush to disk while it was running, in fact I only ever saw it happen once the CPAP machine was stopped. So I thought I would seek the advice of people here. Is there some kind of cable I can buy to fit the proprietary hardware slot that normally a modem would fit into and maybe pull metrics that way? (Btw all my work is open sourced, so nothing proprietary!).
Anything helps, thanks!
My machine: Philips Respironics REMstar Auto A-Flex (System one), pictures included in post.
The longer version
My name is Dan, a Software Engineer that has recently started taking an interest in having more ownership over my sleep apnea, namely at the software/hardware level. I've was diagnosed with sleep apnea probably 12 years ago. A big turning point for me was finding SleepyHead and realizing just how much information and personal sleep data was being withheld from me. I've read over the source code for SleepyHead and learned a ton from it, I even ported a custom data parser specifically for my machine type. However I felt like there was a next step missing. I wanted my sleep data in real time, there is a lot you could do with that information. What if you had custom machine learning algorithms that took in all of you historical sleep data and could identify predictors for when you'll have an apnea event? You could then use the streaming metrics to adjust certain therapy characteristics (or real world stimulus) in real time for better patient outcome. My current goal is a lot smaller then that though, I actually just want to use it for a creative fun artsy project to raise awareness for sleep apnea. If you haven't already read the TLDR above I've basically hit a wall and have almost given up. However I recently read a VICE article (can't link, new member) about SleepyHead and Mark Watkins that has really fired me up again. There has to be a way to retrieve these metrics in real time. It's done during sleep studies and my hope is that there is a cable for my machine that allows direct access to some internal memory or buffer where I can consume these metrics as the machine processes them. If you know anything please let me know!
