Interesting. I guess what I was getting at is that I have intimate knowledge with this problem, for the elderly and the same application of what your brother is

Interesting. I guess what I was getting at is that I have intimate knowledge with this problem, for the elderly and the same application of what your brother is trying to do. We wanted to know how they lived, if we could detect and predict anomalies (heart, falling, etc.) We worked on a similar project: interfaced with a fitness tracker through bluetooth. Wrote the code to decode bluetooth packets and send commands. Put the code in raspberry-pis connected with a 3G modem, auto-updated, auto-paired with devices, with the actor system. Data went through Kafka. The back-end was in Scala. We had alerts. We had an Android app. We had dashboards in Grafana. A lot of thought went data upload schedules and what to send not to max the data plan, etc. The whole shebang, thousands and thousands of commits. I was actually hired to the company to design and build that product, interface with the hardware to get the data, etc. We also wanted to detect falling, and thought about several ways to do that. I set to go directly to the accelerometer's data, and tried to study the change in RSSI with respect to several "base stations" on different planes. Ask me how many people use it? We will use the code base for other applications that require streaming data from unattended devices that auto-update and come back to life on their own when they fail.