Brontoligist, the lightnig tracker¶
Brontoogist is an open-source lightning tracker with and emphasis on open data and collaboration. The system is designed around observer nodes that can participate in one or several networks of lightning trackers as well as locally store observations.
The goal and philosophy is build a system where anyone interested in "electromagnetic" weather can contribute, either by hosting observer nodes, writing server side real time lightning localization algorithms or just chime in to the discussions.
Observer¶
The observer continuously records the signal from one or several antennas tuned to collect lightning strike signals. When a signal with an amplitude high enough is detected the system will store a lightning record. This record contains a little data before the trigger, some longer data after the trigger and the precise time of the trigger. Time is tracked with GPS to ensure that each node in a network runs on the same clock and recordings can be correlated later. Recording and precise time keeping is done with a STM32 MCU that sends each record to a raspberry pi or other computer for further analysis.
The raspberry pi or other computer can then further analyze the record to early reject noise and interference, the record is then passed on to one or several networks or local storage.
Networks¶
Networks are collections of observer nodes with a shared central server to where all the records are sent. The server side software is open-source and build to make it easy to experiment with various algorithms to analyze the stream of lightning records.
Contributing¶
If you are interested in lightning or "electromagnetic weather" in general, feel free to contribute to the project. As a sole initial developer with an electronics background there are many areas where help would really help.
AI¶
This is a spare time project spanning multiple disciplines started by a solo developer, AI coding tools will be used to keep the speed up during development. The idea is not to have the system to be fully AI slop but to be able to get to a proof of concept system in a reasonable amount of work. Also, large parts of the system is quite non-unique IoT data ingestion chain, hopefully with enough examples for the AI to have learned well enough.
Human contributors interested in the goal of the project are more than welcome to replace the AI in this project. :)