MLAT speeds highly improbable

[andrewsi]

I took a look at the track logs from two flights I made earlier today in N688PC (rented SR-20, KPAE to KFHR and back).

In the track log, the data points from ADS-B all appear fairly consistent with the flight, air speeds rising through 118-130 in the climb and settling at about 140kt in cruise. However, during several portions of this flight the data switches over to MLAT and then the speeds are just all over the place, 174 in one reading, dropping to 95 in one reading and 241 in the very next data point. The graph is crazy spiky and wildly inaccurate even as positional tracking remains fine.

Is this a known issue or something unique to this particular area? Or airplane? Just curious.

[obj]

Yes, mlat speed data is not very accurate, especially if the aircraft is doing anything other than flying in a straight line.

[petersk1]

I had a similar experience with a flight yesterday… speeds showing up to 900 MPH with an actual flight speed of around 165 MPH. That was in a straight line.

https://flightaware.com/live/flight/N53194/history/20210402/1211Z/KOZW/68IS/tracklog

When the flight was being reported by Chicago Center the speeds were fine.

What’s going on there? Differences in timeclocks between reporting sites?

[obj]

Reported mlat speeds are essentially the velocity component of a constant-velocity-model Kalman filter. The input timing data is very noisy so the velocities can jump around a lot, especially when there are few receivers contributing data.