Originally Posted by
Airbubba
And, if some packets are missed, you do get a few herky jerky turns in areas of poor coverage as the FR24 constant velocity Kalman filter attempts to connect the dots.
No argument there, although I wasn't aware that FR24 used any kind of estimating algorithm other than crude extrapolation and dumb joining-the-dots.
That's why a track on the map is only useful given a sufficient plot density and, preferably, timestamps for each point.
Ironically, given the current context, the worst example I've seen was where FR24 turned a lazy 270 into an instantaneous 90 in the opposite direction ...