A custom wrist device solution offers the possibility for significant improvements in terms of reliability and suitability for cryonics monitoring compared to using off-the-shelf fitness trackers. Note: This project has been fully scoped in terms of cost and timelines with hardware specialists and requires funding in order to move any further.
A custom built wrist device has many advantages compared to commercially available devices. A photoplethysmogram (PPG) sensor optically measures blood volume changes via shining light into human skin. This sensor, in varying forms, is used for passive measurement of heart rate, blood pressure and blood oxygen on the wrist in commercial devices. A custom device offers the possibility to read raw data from the PPG sensor which is very useful for developing a better cryonics monitoring solution as it allows us to tailor the processing of the data, using machine learning methods, towards prioritizing accuracy in detecting the absence of vital signs. We estimate that by analyzing raw PPG sensor data the device can reliably detect (with as much as 90% accuracy) that the wearer has no pulse within less than a minute.
A custom device can also have raw and continuous accelerometer access, giving it a significant edge over the method of creating apps on existing consumer devices or simply looking at health app data such as Apple Health or Google Fit. After many years of experimentation and prototyping we have determined that you cannot reliably get runtime and access to sensors on existing commercial devices, and this is a critical component to having reliable monitoring. The Fitbit Sense was an exception which is why this was a device of choice when it was available.
Our strong belief at Cryonics Monitoring is that other vital sign monitoring be 'backed up' by analyzing motion data - i.e. have you moved at all in the past hour. Since other vital sign data has a non-trivial possibility of false negative readings, we will always sound the alarm if you don't move after some of amount of time, making the prospect of false negatives in our monitoring paradigm slim to none.
Another key advantage of a custom device solution for cryonics monitoring is that there will not be any software updates which are not thoroughly tested regarding their effect on death detection time. Commercially available devices such as the Apple Watch, Samsung Galaxy and Fitbit devices frequently have automatic updates by the device manufacturer. Any of these updates could potentially change something regarding the sensor data preprocessing and make the implemented death recognition program less reliable or even completely non-function.
Through experimentation we observed that two different software versions of the Apple Watch (version 7.1 vs 7.5) derived different heart rate numbers on an inanimate object, drawing attention to the fact that even minor version updates could modify sensor processing in ways that alter vital sign data. In order to maintain accuracy in detecting the absence of vital signs we must be able to control firmware and sensor processing updates.
What would be the components of a custom device? The most noteworthy components / hardware features that the custom device will have are the following:
Below is a high-precision accelerometer. A part like this is typically under $10, even when purchased in small volumes.
The PPG sensor will be used as the main way to detect emergencies. The accelerometer and the gyroscope will detect motion and be used as a fallback death/emergency recognition. Whereas our prior cryonics alarm solution (development stopped due to Fitbit Sense being discontinued) required a Bluetooth connection to a smartphone with internet access to send alarm notifications, a wrist device built from scratch by us will have internet access on its own via Wifi and/or a LTE transceiver. Wrist detection avoids false alarms when the user temporary does not wear the device. All components (in total) of the custom device are estimated to cost less than $80. Thus, at cost, the custom device will be cheaper than almost any smart watch currently sold.
Popular commercially available smart watches, like the Fitbit Sense, run proprietary power management software programs to achieve longer battery life during normal usage. Implementing equally efficient power management programs will probably not be achievable within the scope of this project, however, measures will be taken to preserve battery where possible. Since our device will need neither high computation power nor a big screen, the power consumption will nevertheless be much lower than that of most smart watches. We aim to maintain a 3+ day target for battery life.
A custom wrist device solution for cryonics monitoring offers several advantages over commercially available devices. It will allow us direct access to sensor data, will not be subject to third party automatic software updates that could affect the reliability, and will have its own internet access which removes the dependence on a smartphone. By tailoring a device like this (wrist-worn sensor device) to the purpose of cryonics monitoring, we can achieve great speed and reliability, with no threat of the device being discontinued or modified by the producer in a way that compromises our use case.
After almost 3 years of prototyping and experimentation, we believe that building from scratch is the only way to achieve an acceptable speed, reliability and usability for cryonics monitoring. It's possible a new consumer device may come on the market that is suitable to cryonics monitoring purposes, but for a number of reasons we see this as unlikely, and people in cryonics have been holding out hope for this prospect for many decades, with much disappointment. After years of research and experimentation, we know how to build from scratch, including which sensors are critical and what kind of OS behavior is required.
However, without substantial investment into cryonics monitoring, such a project cannot be completed by our team, as in order to get optimal battery life, smooth performance and reasonable development timelines, the hardware and firmware components would be best overseen by a specialist in these things, which none of our current team members are.
Stay tuned for how you can support us in creating a system like this!