Is it possible to implement a machine learning adaptative data rate algorithm ? It will be used if we have a mobile node. So, the algorithm would need training based on the GPS data sent by the device in order to predict its next position and the data rate acoordingly.
Is it possible to implement a custom machine learning adaptative data rate algorithm by adapting the default.go file?
Chirpstack - opensource software, anything is possible.
The only problem I’m thinking of is where to store the gps data being sent and how to integrate it to the chirpstack compiled version. I guess I didn’t state it clearly. Sorry.
Please note that the NS does support ADR plugins. That way, you can implement your own ADR algorithm, without the need to make modifications to the ChirpStack code. See: chirpstack-network-server/main.go at master · brocaar/chirpstack-network-server · GitHub.
You can compile that to a Go binary and configure it under
adr_plugins in the TOML config (path to the binary).