Accuracy of Tracks From a Garmin Etrex GPS Receiver: Saved Tracks Vs. Tracks in Memory
GPS has become interesting for hiking now that the accuracy degradation (SA) of the GPS signals has been turned off. The accuracy of GPS position determinations is now routinely quoted as ~15-40'.
Unfortunately, the manual for the Garmin Etrex GPS Receiver gives no information at all about the tracks stored in memory or the tracks stored in saved tracks. Hence I have begun to analyze the accuracy of those saved tracks. The Etrex software version used for this analysis is 2.05. (Note that there are free later software updates.)
Here are some characteristics of the saved tracks, most of which were a surprise to me:
- The Etrex does not store the individual points which are acquired every second in normal operation and every five seconds in battery save mode.
- Instead, the Etrex does something more clever. It filters the individual points to produce a more accurate determination of one's track, and it then approximates one's track as a series of straight lines. This has two advantages: it removes some of the jitter caused by position errors, and significantly cuts down on the storage required for a given track.
However, this is tricky software. If one is travelling only in a straight line, the software is easy and works wonderfully. However, if one is switchbacking, the software has to determine at what point one has stopped going in a straight line and begun a different straight line. Overall, my initial impression is that Garmin seems to have done a pretty good job.
- The Etrex can store ~1200 points in the memory. The storage includes latitude, longitude, elevation, date, local time, and whether a new track was started after the loss of satellite signal or after the Etrex was turned off. Unfortunately, the estimated error of the position determination is not saved!
- Saved tracks use further software that approximates the track stored in memory with a lower quality rendition in order to reduce the number of points that are saved. Further, the time information is discarded.
- The accuracy of saved tracks from a Garmin Etrex can vary tremendously. If the memory includes only a small region with few points, the saved track can be a pretty faithful rendition of the track as stored in memory. However, if the memory includes a large region with many points, the saved track has much lower accuracy than the tracks saved in memory.
Hence unless you don't mind losing accuracy and time information, I recommend that you never save the memory as a track. I suspect, but haven't verified, that simply doing the save operation cuts the number of points in the memory as well.
Here is an example of the accuracy degradation for a Grizzly Flats hike in the San Gabriel Mountains. I used the Etrex in normal mode to record a 100 mile journey from my home in San Diego County to the Upper Dark Canyon Trailhead at mile marker 30.02 on the Angeles Crest Highway. I then used the Etrex in battery save mode to record the hike. (Normal mode is necessary to obtain accurate tracks for car travel, but battery save mode is sufficient for low-speed travel.) In the hike, we traveled to the saddle above Grizzly Flats, but then took an excursion along the Fire Road to Mt. Lukens before returning to the saddle and going on to Grizzly Flats.
I downloaded the memory without doing any saving of tracks. I then saved the memory as a track, and downloaded the saved track. The 1440 points from the original track was compressed to 125 points in the saved track!
A comparison of the two tracks for the "to" part of the trip (the "return" part is not shown for clarity) shows that the saved track is an extremely poor approximation of the original track stored in memory. There are two places along the track where the saved track is off by ~650'!
Note that this large error may be due to this being a portion of a track with a 100 mile journey. I have seen other examples where the saved track is an excellent approximation to the track in memory, when the entire track in memory spanned only a few miles.
See the topo map for a comparison of either track to truth. Note that even the track in memory did not record the switchback shown on the topo map.
Copyright © 2000-2001 by Tom Chester.
Permission is freely granted to reproduce any or all of this page as long as credit is given to me at this source:
Comments and feedback: Tom Chester
Last update: 13 July 2000 (minor updates 2 January 2001).