GPS Questions Part 2

[DayTrip]

I bought a GPSMAP 62s about a month ago now and have some follow up questions to my original thread.

1) I noticed that the Trip Odometer mileage rarely agrees with the actual mileage I walk. It is much farther, at least for the 4 hikes I have tracked and followed. When I load the tracks in Base Camp and look at the Track Details it is normally pretty close though. As an example, I did Kinsmans this past Sunday and AMC Guide mileage of trails I took was 11.5 miles and I think Track stats showed 11.3 miles. The Trip Odometer distance of the walk was over 13 miles (I usually do an "ascent" track and a "descent" track so I am adding the 2 Trip Odometer distances and I always reset before I start officially walking). ??? Is it because the Trip Odometer is live and measuring every back and forth step, zig-zag, etc where the Track is taking readings at certain intervals? I'm still using the GPS with "out of the box" settings so I don't know if I should be tweaking something or if they are both accurate given when they are measuring. I did a lot of zig-zags along this particular hike but I wouldn't think it was 2 miles worth but maybe it was. Another thing I wonder about is how much distance is captured when you are not moving and the GPS signal is fluctuating and creating "distance moved" that is really not walking distance. You can see squiggles in the track but it does not appear to effect Track stats but it does affect Trip Odometer.

2) Is there an easy way to sync tracks and way points between Base Camp and device? I generally mark way points by using the default number assigned and then rename in Base Camp after when I upload based on what it was (trail jct, river crossing,etc). It appears renaming the way points creates a second "duplicate" with the original way point so you have to manually delete the numbered way points, which gets confusing of there are a lot of way points. If I then try to sync with device it has both sets, which is very annoying. Also, on the GPS is there a way to select way points just from a given track or does it always have all the way points you load? This is also very annoying when you are hiking in a similar area to another hike. Is there a "best practice" to manage this? It is not an issue in Base Camp because you can pick the track and sort but I don't see anyway on GPS to do this. I don't see an archive option on GPS for way points but there is for the Track.

If anyone has feedback on this it would be appreciated as always. Thanks.

 

[Guthook]

I can take a stab at question 1, since I've been making tracks of the AT for my iPhone apps with a Garmin GPS. I get a similar difference between the trip odometer and track mileage as well, which I think is exactly for the reason you say– the track cuts corners sometimes, while the trip odometer records constantly. When you say "actual mileage," though, that assumes the guidebook has the exact correct mileage. I don't know about the White Mountain guide, but the Appalachian Trail's official mileage is often pretty far off from what is "actual". Especially since nobody runs a measuring wheel all the way up the AT every time there's a relocation somewhere– if I understand right, they measure the relocation and calculate the change in mileage from the old section. After doing this a few dozen times over the years, there's bound to be some rounding errors somewhere. My guess is that the "track mileage" is the closest to what is actual, since it can be objectively measured. How exactly is every trail in the WMG measured, and how often are routes checked for accuracy? Every measuring method will produce a slightly different number because they're all measured slightly differently.

 

[akafuzzjones]

There is evidence of some of these differences when the mileage on a new map doesn't match the mileage in an older guide book. Then in some instances the even older trail signs have a third milage that is different from the other two. Have seen this with some of the AMC Maine guides and maps.

 

[DayTrip]

I caved in and upgraded my 1986 AMC Guide and maps this year with the latest edition, which claims to have remeasured with GPS, so I am assuming the mileages in the new stuff is reasonably accurate. Thanks for the feedback. Was wondering if it was settings on my GPS or just the increased precision of the GPS. It is amazing how much more mileage you put on relative to the "official" distance. Like golf I guess. I know with my game I'm walking well over 400 yds on a 400 yd Par 4. Guess I'm the same way hiking. I'll go with my Track distance and stats as "official statistics".

 

[DougPaul]

1) I noticed that the Trip Odometer mileage rarely agrees with the actual mileage I walk. It is much farther, at least for the 4 hikes I have tracked and followed. When I load the tracks in Base Camp and look at the Track Details it is normally pretty close though. As an example, I did Kinsmans this past Sunday and AMC Guide mileage of trails I took was 11.5 miles and I think Track stats showed 11.3 miles. The Trip Odometer distance of the walk was over 13 miles (I usually do an "ascent" track and a "descent" track so I am adding the 2 Trip Odometer distances and I always reset before I start officially walking). ??? Is it because the Trip Odometer is live and measuring every back and forth step, zig-zag, etc where the Track is taking readings at certain intervals? I'm still using the GPS with "out of the box" settings so I don't know if I should be tweaking something or if they are both accurate given when they are measuring. I did a lot of zig-zags along this particular hike but I wouldn't think it was 2 miles worth but maybe it was. Another thing I wonder about is how much distance is captured when you are not moving and the GPS signal is fluctuating and creating "distance moved" that is really not walking distance. You can see squiggles in the track but it does not appear to effect Track stats but it does affect Trip Odometer.

This question has been asked and answered on this BBS many times...

The length of a trail is a fractal--different methods of measuring the length give different distances. There is no fundamentally correct length of a trail. Furthermore, all measurements contain errors. Search on the word "fractal" for more info.

Doug

 

[DayTrip]

Sorry if I'm treading on "old news". New to GPS and never heard of fractals. I'll research.

 

[DougPaul]

New to GPS and never heard of fractals.

I think many (if not most) hikers are unaware that trail length is a fractal. It just happens that the word has been used in recent replies to the question and therefore searching for it is a good way to find relevant threads on this BBS.

It may also be worth noting that your GPS is more accurate than many maps (actually both have errors...) and they will often disagree, sometimes by surprising amounts. (Another frequent beginner question...)

Doug

 

[Remix]

I think many (if not most) hikers are unaware that trail length is a fractal. It just happens that the word has been used in recent replies to the question and therefore searching for it is a good way to find relevant threads on this BBS.

It may also be worth noting that your GPS is more accurate than many maps (actually both have errors...) and they will often disagree, sometimes by surprising amounts. (Another frequent beginner question...)

Doug

There are many ways an odometer reading can be computed....by summing the distance between every 1 to whatever number of computed position locations. Each computed distance location may have alot of error when hiking (varying signal coverage and constantly switching between sets of satellites). Looks like the two are working together to give you bad odometer readings.

Its not uncommon for the odometer reading on hiking gps's to have a large error.

Nontheless my car GPS readings are usually dead on, even though roads are "fractal". So its a combination of poor software development and noisy position reports. Since you have a mapping unit, the odometer really isn't necessary.

If you really want to experiment, try changing how often the gps records position and see if accuracy changes.....I would do that in a car or on a bike .

 

[DougPaul]

There are many ways an odometer reading can be computed....by summing the distance between every 1 to whatever number of computed position locations. Each computed distance location may have alot of error when hiking (varying signal coverage and constantly switching between sets of satellites). Looks like the two are working together to give you bad odometer readings.

Consumer hiking GPSes compute one position per second and a track is some subset of these positions. The trip computer uses the one per second data. I believe the total distance is computed as the sum of the distance between successive positions. (If the distances are computed in this way, the track length must be less than or equal to the trip computer distance.)

We reported an experiment where we analyzed the repeatability of GPS tracks on a treed trail in a narrow valley (S Greely Pond Tr). This experiment showed that modern high-sensitivity GPSes are actually pretty accurate when used properly in realistic conditions--see http://www.vftt.org/forums/showthread.php?t=14406 for details. (Short-term error changes have a far greater effect on the computed distance than long-term or absolute errors. (A constant error will have no effect on the computed distance.) I have (unpublished) data which shows the short-term error change to be much smaller than the long-term or absolute error.)

Nontheless my car GPS readings are usually dead on, even though roads are "fractal". So its a combination of poor software development and noisy position reports. Since you have a mapping unit, the odometer really isn't necessary.

Paved roads are smooth (as seen by the car body) and have very gradual turns (on the scale of one position per second) so the fractal dimension is essentially 1. A fractal dimension of 1 means that the GPS speed and your car speedometer should give the same reading...

BTW 1: Car speedometers (and odometers) can easily be off by up to several percent because they don't know the exact circumference of the tires (which varies with air pressure, model, and wear).

BTW 2: if you travel at a steady speed on a reasonably straight section of road, the GPS (rated accuracy .1mph) is more accurate than your speedometer...

If you really want to experiment, try changing how often the gps records position and see if accuracy changes.....I would do that in a car or on a bike .

Better experiment: record a 1/second track and compare its length to the length of sub-samplings (of the same track). It is trivial to prove that the lengths of the sub-sampled tracks are less than or equal to the length of the 1/second track. (Proof by triangle inequality.)

Doug

 

[Trail Bandit]

I have found that the accuracy of the GPS track is often reduced by poor satellite reception. The newer GPS units from Garmin have better antennas and something called a SiRF receiver. I currently use a GPSmap76Cx unit which is far better than the GPSmap76S that I started with. That said, I realize that my current model is now "OLD". Where you carry the GPS receiver makes a difference too. My unit comes with a carrying case that has a belt loop so you can carry it on your waist, but in doing so your body blocks the signal from some satellites and that will change as you turn in different directions. If I am collecting tracks for a map, I use a remotely mounted antenna that has a cord to plug into the receiver. I have a baseball cap with a tin can cover attached to the top and put the magnetic antenna on it.
This setup produces more consistent tracks with less scatter and better reproducibility than just the receiver on the belt approach. For a more accurate trail location for a map I try to hike the trail a number of times and take an average of the different tracks. For the scale that is used for most paper maps it makes no difference but if you give out track files for some obscure trail for others to follow it is better. When you zoom way in on any GPS track, the points are scattered back and forth along your route. By averaging a number of tracks and smoothing the result you get a more accurate result and it doesn't have all those zigs and zags that really aren't there. My GPSr computes the calculated accuracy of the position it is reporting. The best I have seen is +/- 9 feet, with the unit sitting vertically on top of a post in an open field on a nice day. My receiver had the WAAS (wide area augmentation system) turned on for that test. Most times, the reported accuracy is worse, sometimes a lot worse. That said, even if the track you are following perfectly follows the actual center line of the trail, your receiver may show you off the trail if you zoom way in.
On the issue of the difference in trip distances reported by the GPS receivers I am still confused. My friend Mary and I have identical receivers, in identical belt cases (neither using the remote antenna) walking over the same route together and her unit consistently records a shorter trip length, odometer, etc. To the best that we can figure out, the settings on both units are the same. The only factor that I can think of is that usually I walk first with Mary following at the appropriate distance behind and in doing so, I have used up some of the GPS signal as I go by, thus leaving her receiver with a lower signal strength.

 

[Peakbagr]

Following advice from Doug Paul and others, and from another friend, I've experimented with varying track point recording methods on my 60CSx. I used to record on a time basis and for the last couple of months migrated to a distance basis. 0.01mi, 0.02mi, 0.04mi, 0.05, etc.
Using the more frequent recording times for the tracks( 0.01mi, 0.02mi ), the track log shows much more detail for a hike or whack.
But what's consistent over the 5-6 years of using the unit (average of 1-2 hikes a week) is that the mileage and elevation shown at the end of hikes is always higher, sometimes considerably higher than when the track is downloaded to the computer.
The GPS numbers are usually higher than guidebook numbers as well.

 

[DougPaul]

I have found that the accuracy of the GPS track is often reduced by poor satellite reception. The newer GPS units from Garmin have better antennas and something called a SiRF receiver. I currently use a GPSmap76Cx unit which is far better than the GPSmap76S that I started with. That said, I realize that my current model is now "OLD". Where you carry the GPS receiver makes a difference too. My unit comes with a carrying case that has a belt loop so you can carry it on your waist, but in doing so your body blocks the signal from some satellites and that will change as you turn in different directions. If I am collecting tracks for a map, I use a remotely mounted antenna that has a cord to plug into the receiver. I have a baseball cap with a tin can cover attached to the top and put the magnetic antenna on it.

The 76Cx (also 60Cx, 60CSx, and 76CSx) were the first models with the SiRF III (modern) high-sensitivity receiver chipset. (Much better than the previous units in less than perfect signal conditions.) The 76 (and 78) series use patch antennas and the best orientation is horizontal (as when sitting on a chart table). The 60 (and 62) series use a quad helix antenna and the best orientation is vertical. (In both cases, the axis of the antenna is vertical.) The improvement over the previous models is due to the high-sensitivity receiver, not the antennas. (The same antennas were used on previous units.) The more recent lines (62, 78, and others) use different high-sensitivity receiver chipsets but are reported to have similar performance to the 60x and 72x series. ("x" in the 60 and 76 series denotes SiRF-III high-sensitivity receiver.)

How to carry a GPS was covered by my phrase "when used properly". A properly oriented antenna (eg a "flat up" patch antenna or a vertically oriented quad helix) in a clear location (eg on top of one's head) works better than other locations or orientations. Properly oriented in the top of your pack is practical and generally works well for general hiking. (If I am mapping a trail, then I use an external antenna mounted on the top of my hat. Professionals generally use an external choke-ring antenna on a pole.) If you carry your GPS on your belt and complain to me about the accuracy my response will be "don't waste my time...".

My GPSr computes the calculated accuracy of the position it is reporting. The best I have seen is +/- 9 feet, with the unit sitting vertically on top of a post in an open field on a nice day. My receiver had the WAAS (wide area augmentation system) turned on for that test. Most times, the reported accuracy is worse, sometimes a lot worse. That said, even if the track you are following perfectly follows the actual center line of the trail, your receiver may show you off the trail if you zoom way in.

The accuracy reported by the GPS neglects sources of degradation (eg multipath) and therefore means little--it is just a lower bound on the accuracy (ie the error is greater than or equal to the reported number). (Note: GPS accuracy is a statement of error probability: the error will be less than the stated distance some percentage (typ 90 or 95%) of the time.)

I don't use WAAS for hiking--the WAAS satellites are low on the horizon and easily lost. The track jumps whenever WAAS is acquired or lost. (It is (intended for) and fine in an airplane where the WAAS satellites will not be blocked.)

On the issue of the difference in trip distances reported by the GPS receivers I am still confused. My friend Mary and I have identical receivers, in identical belt cases (neither using the remote antenna) walking over the same route together and her unit consistently records a shorter trip length, odometer, etc. To the best that we can figure out, the settings on both units are the same. The only factor that I can think of is that usually I walk first with Mary following at the appropriate distance behind and in doing so, I have used up some of the GPS signal as I go by, thus leaving her receiver with a lower signal strength.

I have done tests where I carried two identical GPSes side-by-side on an out-and-back and compared the outgoing and return tracks--sometimes one GPS will do better, sometimes the other*. Whatever you do, there will always be some systematic** and random variation. Be happy if your reported distances are within a few percent of each other.

* The comparison would have been better if I had fed both GPSes from the same external antenna, but I don't have the hardware at hand.

** No two units are truly identical--there will always be some variation in the parts.

Doug

 

[DougPaul]

Following advice from Doug Paul and others, and from another friend, I've experimented with varying track point recording methods on my 60CSx. I used to record on a time basis and for the last couple of months migrated to a distance basis. 0.01mi, 0.02mi, 0.04mi, 0.05, etc.
Using the more frequent recording times for the tracks( 0.01mi, 0.02mi ), the track log shows much more detail for a hike or whack.

When I am performing bakeoff-style experiments, I record track points strictly by time. (Any recording scheme that uses position (eg distance or auto) risks biasing the results.)

When hiking (or driving), I just use "auto" "norm". (This records points more often when I am moving slowly or turning a lot.) I'm not obsessed with recording every footstep... A recent hike (the Kinsmans) was 11.1mi by GPS and 11.3mi by the guidebook--close enough for me.

Other users, of course, are free to choose whatever they want.

But what's consistent over the 5-6 years of using the unit (average of 1-2 hikes a week) is that the mileage and elevation shown at the end of hikes is always higher, sometimes considerably higher than when the track is downloaded to the computer.

The GPS and your computer are free to process the data any way they want in the process of computing the track length. I suspect that some smoothing may be applied (particularly to the altitude data) in some cases. Of course, the GPS always has access to the one-per-second position data while your computer only has access to the subset of the data that is recorded in the track.

Doug

 

[Osmosis]

I use the Garmin Fenix wristwatch. Even though there may be discrepancies between GPS and maps the real benefit for me is knowing distance hiked vs. distance to go.

 

[bikehikeskifish]

Over the course of a 100-mile bike ride, the bicycle computer (calibrated via a roll out) and the GPS (76CSx in my jersey pocket (yes, Doug, I know this is not ideal)) are at most out of agreement by 1/10th of a mile. Good enough for me

When hiking, distance almost always comes up shorter than the guide and elevation almost always comes up greater.

Tim

 

[Peakbagr]

Tim,
My result is that when hiking, GPS distance almost always comes up longer, and elevation comes up greater than when the track is downloaded.

 

[DougPaul]

Over the course of a 100-mile bike ride, the bicycle computer (calibrated via a roll out) and the GPS (76CSx in my jersey pocket (yes, Doug, I know this is not ideal)) are at most out of agreement by 1/10th of a mile. Good enough for me

A jersey pocket while biking is better than on your belt while hiking--the GPS has a better skyview due to the cyclist's forward lean. (I have cycled with an external antenna on my helmet when performing some experiments...)

IIRC, you have a 76CSx or 76Cx--the antenna orientation will be close to optimum when your chin is almost on the bars which also minimizes the wind resistance...

There is no problem with carrying the GPS in sub-optimal locations as long as one is aware of the trade-offs and has realistic expectations.

Doug

 

[bikehikeskifish]

Mine too... I guess I was unclear. The reported total distance is shorter than the WMG. The reported total elevation is greater than the WMG. The tracks, when downloaded, are usually shorter in both dimensions than the total reported. It depends on where you download the tracks to. Google Earth reports different values than Garmin Mapsource. MapMyRide.com varies too, although I haven't examined that one as closely, and I don't usually bother with Garmin Mapsource for bike rides.

Tim

 

[DougPaul]

Mine too... I guess I was unclear. The reported total distance is shorter than the WMG. The reported total elevation is greater than the WMG. The tracks, when downloaded, are usually shorter in both dimensions than the total reported. It depends on where you download the tracks to. Google Earth reports different values than Garmin Mapsource. MapMyRide.com varies too, although I haven't examined that one as closely, and I don't usually bother with Garmin Mapsource for bike rides.

Garmin MapSource and National Geographic also yield different distances and elevation gains from the same track file.

I wrote a program to compute the total distance and elevation gain from a track file--it gave yet another set of results for a typical hike track but at least I knew the processing that was used... Track elevation was particularly noisy and some sort of smoothing had to be applied to reduce the noise. (In practice, as I increased the smoothing I also reduced the computed elevation gain. So by adjusting the amount of smoothing, I could get any elevation gain I wanted...)

The characteristics of a GPS are known to its internal smoothing and it can be optimized to match them. There is also information (eg satellite constellation and signal strengths which can be used to help estimate the horizontal and vertical errors moment by moment) known to the GPS which may also be used. This information is not placed in the track file so an external program cannot use them. My track files do not even identify the GPS model so the program cannot apply any GPS-specific processing.

Just a few more reasons why different programs often produce different results...

Doug

 

[Trail Bandit]

This is a bit off topic but it is a GPS reading that you don't see very often. The waypoint was captured at the location shown but the photo was taken later in camp. Note: the Garmin GPSmap 76CSX will not permit manually loading in the N90 location, I tried. It will allow 89Deg, 59 Min 59.9 Sec. To get to 90 Deg North, you actually have to go there. They are great hikers who I met on St. John, VI and wonderful people. I hope to be at least 1/2 as active as they are when/if I reach their age.
The following is from the Guinness Book of World Records.

The oldest married couple to visit the North Pole are Heinz G.Fischer (b.10 March 1929) and his wife Linda G. Burdet (b.29 November 1931) (both USA) who skied to the North Pole aged 79 and 76 respectively on 12 April 2008.

The couple were dropped by helicpoter and skied 6.39 km to ice camp Barneo (1 degree south of the pole) where they camped overnight. This was part of an expedition but the expedition leaders at the time had never taken couples over 40 to the ice camp and it was they who suggested the couple submit a claim.