GPS accuracy experiences

Ask questions and share your experiences with hiking & backpacking gear, and share trail recipes and gadget tips. Please see classifieds forum for buying/selling stuff.
johnspeth
Posts: 156
Joined: July 30th, 2013, 8:33 am

GPS accuracy experiences

Post by johnspeth » October 10th, 2019, 9:47 am

I've been using my son's running GPS watch (Garmin Forerunner 220) for the past year. I used to think it was reasonably accurate until I hiked with a friend who also used a GPS, his built into his iPhone. We discovered that my Garmin watch measured distances that are about 20% longer than the iPhones'. That's an unacceptably high error for me.

Has anybody reading this experienced with the same problem with their GPS devices? Why does this happen to my watch? What did you do to mitigate the problem (new device?)?

I've characterized the error empirically but now I find myself doing mental calculations in my mind while hiking. It's not the kind of mental effort I like to use while hiking.

User avatar
kepPNW
Posts: 6345
Joined: June 21st, 2012, 9:55 am
Location: Salmon Creek

Re: GPS accuracy experiences

Post by kepPNW » October 10th, 2019, 10:02 am

johnspeth wrote:
October 10th, 2019, 9:47 am
measured distances that are about 20% longer
In my experience, it's easy for two identical devices to be off up to 10% simply due to different sampling rates. In general, the more data points there are the longer the line is. Phones don't tend to record points very often, and Garmin devices (eg, the "trip computer" on their GPS units in particular) recalculate results up to once per second.

Best confidence booster, or not, would be to take the unit(s) in question along a known route. A few loops around a high school track, maybe? (That'll kill *any* idea that phones accurately record elevation gain!) For me, walking the entire Wildwood Trail, which has mileposts every quarter-mile, told me that the data recording settings on my Garmin were damn close.
Karl
Back on the trail, again...

(Photos · PortlandHikers)

User avatar
teachpdx
Posts: 147
Joined: January 21st, 2014, 4:45 pm
Location: Hillsboro, OR

Re: GPS accuracy experiences

Post by teachpdx » October 10th, 2019, 10:40 am

I've had a bit of an opposite experience.

My best friend and I did 50 miles in the Wallowas earlier this summer, he with his Garmin watch and me with my Apple Watch and iPhone. The Wallowa NF has a GIS site that lists all their trails down to the hundredth of a mile (not that it's that accurate, but it must be close). At the end of everything, my watch/phone was within about 2% of all these listed values, and his Garmin consistently recorded 10% less distance than my watch.

When the Apple Watch is in hiking workout mode, it samples location data every 10 seconds and does a great job picking up the ends of switchbacks, whereas a longer duration sampling would potentially miss some of them and shorten the distance.

I verified this earlier this summer around Mt. St. Helens. The path of a hiking workout can be overlaid on a satellite view, and many trails in that area are visible on satellite view since there is no tree cover. Across the Plains of Abraham and Windy Pass, my line didn't vary from the visible trail by more than maybe 30' at most. It was surprisingly accurate.

This was all using a Series 3... I've read that the Series 4 and 5 are even more accurate.

User avatar
Bosterson
Posts: 1982
Joined: May 18th, 2009, 3:17 pm
Location: Portland

Re: GPS accuracy experiences

Post by Bosterson » October 10th, 2019, 10:40 am

As far as I can tell from poking around online, the Forerunner 220 is GPS only; you need a higher-end/newer model to support GLONASS. In contrast, all iPhones for years have supported both GPS and GLONASS, which is why their accuracy is so good. If you are out hiking with your Forerunner, it's possible that it doesn't have as good of a signal and is thus bouncing your track around, creating extra distance via jitter. You should be able to test this by taking tracks on both the Forerunner and an iPhone while hiking, and then compare both tracks somewhere like Caltopo and see if they differ, and if so, where. I also agree with Karl about sampling rates - the more points you take, the "longer" your hike will be.
Will hike off trail for fun.

Aimless
Posts: 1479
Joined: May 28th, 2008, 10:02 pm
Location: Lake Oswego

Re: GPS accuracy experiences

Post by Aimless » October 10th, 2019, 10:47 am

I have noticed that if I leave my dedicated GPS unit (not a smartphone) turned on while I am stationary for a half hour eating my lunch, it will often record an extra few hundredths of a mile of movement I did not make. That only amounts to 150 or 200 feet of variance, but if there is a slight variance in each sampling taken while on the move, I can see where it might add up to something significant over the course of a hiking day. That's using a dedicated unit and I've noticed it when beach walking, where the satellite signals should be extremely easy to capture compared to under heavy tree cover.

User avatar
adamschneider
Posts: 2998
Joined: May 28th, 2008, 10:02 pm
Location: SE Portland
Contact:

Re: GPS accuracy experiences

Post by adamschneider » October 10th, 2019, 10:53 am

kepPNW wrote:
October 10th, 2019, 10:02 am
In my experience, it's easy for two identical devices to be off up to 10% simply due to different sampling rates.
This.

A lot of devices will, by default or via an option, do "auto" recording, where they only record a trackpoint once the unit has moved a certain distance. (And sometimes the distance threshold varies based on your speed.) Whereas others will record a point after a certain time interval, regardless of whether you've moved or not. The latter will, of course, record longer distances.

GPS Visualizer has a "trackpoint distance threshold" that can help clean up tracks that have too many strictly-time-based points: https://www.gpsvisualizer.com/tutorials ... lters.html

squidvicious
Posts: 249
Joined: May 11th, 2015, 8:41 pm
Location: Troutdale

Re: GPS accuracy experiences

Post by squidvicious » October 10th, 2019, 11:38 am

adamschneider wrote:
October 10th, 2019, 10:53 am
kepPNW wrote:
October 10th, 2019, 10:02 am
In my experience, it's easy for two identical devices to be off up to 10% simply due to different sampling rates.
This.
Found this out the hard way. I use Viewranger on my phone, which lets you adjust that in settings. Suddenly I started getting *really* off (low) distances on familiar routes. Tracked it down to those settings, which had somehow curiously reset themselves.

RobFromRedland
Posts: 986
Joined: May 28th, 2008, 10:03 pm

Re: GPS accuracy experiences

Post by RobFromRedland » October 10th, 2019, 3:48 pm

I always use the GPS Track Editor to clean up the "jitter" from my phone. When sitting or moving slowly there are lots of subtle variations in the track which exaggerate the distance. GPS Track Editor does a pretty good job of smoothing those out. (at least in my experience).

http://www.gpstrackeditor.com/
Life is not a journey to the grave with the intention to arrive safely in a pretty and well-preserved body, but rather to skid in broadside, thoroughly used up, totally worn out, and loudly proclaiming: WOW-What a ride!

User avatar
aiwetir
Posts: 507
Joined: December 10th, 2014, 11:54 am
Contact:

Re: GPS accuracy experiences

Post by aiwetir » October 10th, 2019, 9:09 pm

Lots of it is the algorithm used to calculate the distance and elevation gains too. You can just straight measure the track or "smooth" it to remove jitter. One thing to note is that if you use an iPhone and it doesn't detect movement in the accelerometers AND the GPS it stops updating your location and shows you as stationary. This is a double edged sword of course because sometimes you need your position updated while you aren't moving much
- Michael

johnspeth
Posts: 156
Joined: July 30th, 2013, 8:33 am

Re: GPS accuracy experiences

Post by johnspeth » October 11th, 2019, 5:15 am

Thanks for all the excellent perspectives. I think I have a good idea of the problems and possible solutions. I took my watch out for a 400 meter track test (Fowler Middle School, Tigard) and ran two trials once around per trial. The first trial had me wandering around the infield for 20 meters before it got me back on the track (total = 420 meters). The second trial was accurate to within a few feet. There are large trees partially hanging over one end of the track and I presume those trees were causing the communication disturbance.

I am convinced that the iPhones are producing high quality tracks in terms of accuracy. During my comparison experiments, we were following an alltrails track. My friends' track always matched the alltrails track for distance while mine was always longer.

I'll summarize what I learned:
- Large trees are the predominate reason for GPS inaccuracy. That problem is not avoidable. It's most evident to me when I'm stopped. Once I walked a quarter mile by sitting on a rock for lunch for about a half hour.
- iPhones use Russian GLONASS as well as US GPS. That (and maybe better on-the-fly filtering) probably explains why the iPhone is more accurate than my Garmin.
- GPS Track Editor does a nice job of removing the GPS wander errors. My prototype hike was Hunts Cove to which I stuck to the trail and did little off trail exploring. My Garmin measured 14.2 miles but should be closer to 12 miles according to mileage reports I found. GPS Track Editor got it down to 12.8 miles by running the Local Inconsistencies filter. I reviewed the filtered points and they made sense to me.
- It might be possible to use the GPS device auto pause feature to filter out GPS wander errors while not moving. I haven't used that setting yet but I suspect it will come with some trade-offs that I'll need to experiment with.

Here's my Hunts Cove track:
activity_4143125675.gpx
My Hunts Cove track
(531.01 KiB) Downloaded 10 times
Here's my filtered Hunts Cove track:
activity_4143125675_filtered.gpx
My Hunts Cove track filtered by GPS Track Editor
(216.28 KiB) Downloaded 6 times

Post Reply