insfilterAsync tune Help/guidance wrong?

For the insfilterAsync tune function, see Tune insfilterAsync parameters to reduce estimation error - MATLAB tune - MathWorks United Kingdom

The Help Center says the for the sensorData parameter, GPSPosition says that it should be in meters. When I try and use meters (by converting the NED frame using geodetic2lla) I get an error related to the StateCovariance matrix becoming infinite.

When running the example on the tune page, the GPSPosition values are incredibly small values (1E-5) - which look like latitude, longitude values to me. If you put these into geoplot, they are essentially at the GPS centre. The example's groundTruth are around 1.5 values, which looks like metres, as the Help page says.

I think that the advice on the Help Centre page is incorrect for the sensorData, and GPSPosition should be in latitude, longitude, altitude, not in metres as stated. Can anyone confirm this?