Etrex 20/30: "Stopped time"

[T42]

Wondering vaguely what an Etrex counts as "stopped".

E.g. the other day I had a 20-minute coffee stop that it counted fairly accurately, but <trkpt> entries for that bit show speeds of 0.1 to 0.7 kph, with only one point on 0.0.

[frankly frankie]

Well every recorded trackpoint is only an approximation, so it follows that there will usually be a 'distance' between successive trackpoints, and hence a speed.  if the bike is parked under something like a garage canopy at the time, these could be quite large.
0.7kph equates very approximately to a 'movement' of 1m (assuming the trackpoints are written a 5-second intervals) which is in any case well below the GPS threshold of accuracy.
I imagine the only way round this would be to set the GPS to record trackpoints at specified intervals (eg 50 metres), as opposed to the default 'auto' setting, and that would probably suppress the recording of most stationary trackpoints.   I've never tried it, because the random dithering doesn't bother me, and I think the 'auto' setting generally works very well.

[T42]

So if the position is within the margin of error they count it as stopped but still record a speed. That works.

[Ningishzidda]

IIRC, the 'Satellites' screen shows the unit's estimation of distance accuracy.

One would have thought that if two sucsessive data logs were closer than the unit's idea of its accuracy, the software would determine the unit as 'stopped'. If several successive data logs were jittering around within the accuracy value, its well and truly 'stopped'.

One would have also thought that a unit as complex as a Garmin would have a smoothing algorithm which uses a number of datapoints to decide whether the unit is moving or stood still.

[fuaran]

Depends on just which stopped time you mean. ie on the trip computer on the Etrex, or viewing in some software after downloading it to your computer. They may give different times.
On the Etrex it could be sampling at a higher rate, ie every second, even if it only records a track point every 10 seconds.

If uploading to Strava etc, it will use its own calculations, probably with some smoothing. And it may depend on activity type - ie cycling stopped speed should be higher than walking stopped speed.

[T42]

I have my own prog that allows me to set a minimum moving speed.  I usually use 4 kph, which covers mucking about at traffic lights & such.  Etrex trip computer stopped time is usually a little less than my prog yields.

Smoothing is a difficult one to handle. I do use a smoothing algorithm but it has the effect of shifting the points, which falsifies the time of passage and the speed.  I think I need to smooth then interpolate values for the original points.  Can'o'worms.

[frankly frankie]

On the Etrex it could be is sampling at a higher rate, ie every second, even if it only records a track point every 10 seconds.

FTFY

But it makes no difference - written trackpoints are still approximations, whether it's a single 'raw' sample or an average of 10.

[T42]

I watched the satellite page this morning as I rode through forest.  It looked pretty constant wrt n° of satellites and richness of signal, but the error margin was 8 metres.  My apparent speed was jumping about by up to 10 kph according to tree cover.

I haven't got a conventional counter on the bike just now but when I had I often noticed such disparities, together with a difference of around 2% in total distance covered.  One of the reasons I don't use Strava.

[Ningishzidda]

Well let's face it. The speed and distance readout on a GPS SatNav unit is NEVER correct.

The units are to tell us whereabouts we are so we can reference our position on a paper map. Datalogging is an 'add-on' extra the manufacturers thought might keep us chattering amongst ourselves.
They also do a fairly good job of guiding us along roads and across oceans.