-
esp-csi
Applications based on Wi-Fi CSI (Channel state information), such as indoor positioning, human detection
-
ESP32-CSI-Tool
Extract Channel State Information from WiFi-enabled ESP32 Microcontroller. Active and Passive modes available. (https://stevenmhernandez.github.io/ESP32-CSI-Tool/)
-
InfluxDB
Power Real-Time Data Analytics at Scale. Get real-time insights from all types of time series data with InfluxDB. Ingest, query, and analyze billions of data points in real-time with unbounded cardinality.
-
ESP32-gather-channel-state-information-CSI-
Get CSI frames (Channel State Information) with the use of an ESP32 WiFi chip
So ESP-CSI for human presence detection is a new thing with ESP-S3 I gather, so I ordered a couple of boards and with some investigation with https://github.com/espressif/esp-csi/blob/master/examples/console_test/README.md have the console running using two ESP32-S3 modules (my router didn't seem to support the router-ESP model but the csi_send and console_test seem to be working as described on that link with some messing around effort.
https://stevenmhernandez.github.io/ESP32-CSI-Tool/ https://github.com/StevenMHernandez/ESP32-CSI-Tool or this one: https://github.com/jonathanmuller/ESP32-gather-channel-state-information-CSI- I haven't had a chance to fiddle with any of them but it's certainly cool and I'd love to hear what you end up discovering...
https://stevenmhernandez.github.io/ESP32-CSI-Tool/ https://github.com/StevenMHernandez/ESP32-CSI-Tool or this one: https://github.com/jonathanmuller/ESP32-gather-channel-state-information-CSI- I haven't had a chance to fiddle with any of them but it's certainly cool and I'd love to hear what you end up discovering...
Actually, fall detection for aged care is one of the main areas of study for this subject - there are many public papers discussing this topic in a fair bit of detail. Just googled 'wifi csi fall detection github' and found this https://github.com/honey0920/csi_FallDetection. Unfortunately, many of the successful implementations out there are mostly theoretical, and aren't robust to diverse environments and personnel, so you might have to collect a bunch of data to train whichever machine learning model you choose for your particular application. In saying that, falls are probably one of the easier actions to classify since they create a lot of variance in the subcarrier phase and magnitude, so you might have some luck! Hope that helps, I'd be really interested in seeing this subject work out effectively for a practical application so keep me posted!