https://github.com/jewbetcha/openlaunch/blob/be4a36a58096c2dc94ea8f8fa9affe5269eb4e6b/launch_monitor.py#L64
This value needs to be determined either heuristically via measurement, or analytically using Detection Theory (https://www.ele.uri.edu/faculty/kay.html). You can explicitly set a desired probability of false alarm, and set a threshold deterministically using that value. For this application, it makes more sense in my opinion to simply take some data collections and set a threshold manually based on those observations and manual calibration.
https://github.com/jewbetcha/openlaunch/blob/be4a36a58096c2dc94ea8f8fa9affe5269eb4e6b/launch_monitor.py#L64
This value needs to be determined either heuristically via measurement, or analytically using Detection Theory (https://www.ele.uri.edu/faculty/kay.html). You can explicitly set a desired probability of false alarm, and set a threshold deterministically using that value. For this application, it makes more sense in my opinion to simply take some data collections and set a threshold manually based on those observations and manual calibration.