Why Do GPS Distance Results Vary Between devices?

[Lowell k]

I was testing a Fitbit, Apple Watch, and Gaia on an iPhone and they each gave me different distance hiked results on an 8-mile hike. The Apple Watch and Gaia on my iPhone varied by about a .25 mile.

Does anyone know why this happens or which one is the most accurate?

Thanks,

Lowell

[Roger Caffin]

They are all inaccurate to some degree…..

I suspect the difference may be due to differences in sampling intervals. Maybe.

Cheers

[Rex Sanders]

Like Roger, I’ll name some suspects. Some could create bigger differences than others, but all could be at play.

– Interval differences
What Roger said. The path length measured can vary by how often you sample your location with slight variability. For example, imagine all the “jiggles” in a path made up of 1 second dots versus 5 minute dots.

– Timing differences
Even if two devices measure at the same interval, if their timing differs, you’ll get different results, because there is always some variability in computed GNSS locations for many reasons (keep reading). You would probably see this even if you took two identical devices with identical intervals, but started their internal clocks running at slightly different times.

– Satellite view differences
If one device is on your wrist, and another is perched on top of your pack, they’ll “see” a slightly different mix of satellites due to blockage by your body, trees, and landscape. This will affect both the derived location and variability in location.

– Antenna, chip, and software differences
Even with two devices side by side in the wide open, running on the same intervals, starting at precisely the same time, differences in how GNSS signals are received, processed, and calculated can result in location and path length differences.

– Satellite system differences
Devices may be capable (or not) of receiving any combination of GPS, WAAS, Galileo, GLONASS, BeiDou, and even more obscure signals. Each satellite system provides different coverage over time, with different precision. For example, older iPhones listened for far fewer systems than the latest and greatest.

So which of your devices is “correct?” Impossible to say. In some sense, they are all equally right.

The measuring-path-length problem is both fractal, and stochastic (“having a random probability distribution that may be analyzed statistically but may not be predicted precisely”). Think of measuring the length of a rocky coastline with a 1 foot stick versus a 1,000 foot stick – you’ll get different answers (fractal). Now, have the length of each stick change randomly by a small amount over time (stochastic). Very messy.

More here: https://backpackinglight.com/gps-good-bad-ugly

— Rex

[Matthew / BPL]

The watches may be inferring distance based on their accelerometer. I know my Coros Pace does that and I think my Ambit did.

[Brad W]

I have found depending on the day, my iPhone with Gaia and my Garmin Instinct watch will have up to but not over 10% difference with Garmin being more.(example Garmin reports 9.9 miles, Gaia 9.0) Gaia I have found almost always errors on the conservative side. Known distances will usually report under. I usually export gpx file and use this-https://www.gpsvisualizer.com/profile_input with DEM elevation data set to ‘best available’, calculate elevation gain-yes with 5 for gain threshold. I have found this gives me the most accurate totals of both methods of recording.

[DWR D]

To test which is most accurate, use your car odometer to mark a mile or two on the road. Then go back and walk it with all devices on. (assumes your car odometer is accurate) But don’t walk a 5 mile trail that is listed as 5 miles… chances are no one ever walked that trail with a measuring wheel to actually measure it… and taking mileage off a map is notoriously not accurate… even using a planimeter since trails are often not exactly located right on maps and don’t account for all switchbacks or elevating gains and drops.

D

[Mike W]

Recreational GPS receivers aren’t like professional surveyor’s equipment that record super accurate positions.   The coordinate point that you get is “somewhere” within an accuracy circle of 10 metres.     If you sit on a rock for 30 minutes, you will get a spider web pile of tracks showing in your track records, even though you didn’t move.   Your position as calculated by the GPS will constantly shift because of the inaccuracy of the signal, and the GPS will connect those dots to form a track with a distance (even though you never left the rock).   This plus or minus accuracy is going to be different on all GPS units, even if you are moving (same 10 metre accuracy circle), so the track lengths will always vary.   I walk the same trail repeatedly during the year and the track length is always different (even though I’m using the same gps).

[Lowell k]

Thanks all, clarified what was happening.

Lowell

[Author]

Posts