Jump to content

Waypoint Averaging

Recommended Posts

Magellans all support waypoint averaging. Some of the Garmins, I don't know which, support it.


Off topic but highly relevant to the question:

Waypoint averaging (as spoken of in these forums, i.e. autoaveraging for a minute or ten minutes, or a half an hour) is a useless waste of time. Get a few good single fixes, and if they are close (a couple thousandths of a minute), go with them. The only indication you really have of a good fix is a low EPE. Even these could be quite a ways off, well outside the EPE circle. But you don't have any way of knowing that without much more elaborate procedures, or outside input like high resolutions satellite photos.

Link to comment

Definately average over multiple days...


Guess what, my new 60CS finds all my caches hidden originally w/my basic non-WAAS eTrex to be on the mark coordinate-wise and I've never had a finder complain in over two years.


They were all averaged from readings taken on different days.


The idea is to find the mean. Since an accuracy graph would look like a bell-curve... Or better yet, this real-world example from a beach event where a dozen folks had 10 sats down to the horizon:



(click here to enlarge)


Guess what? The center dot was from a basic non-WAAS non-averaging eTrex.


Averaging won't do much when you have a great constellation. (3+3+3+3+3/5 = 3 obviously)


But when it's bad, it'll give you far better results. (1+5+2+4+3/5 = 3)


You can always tell when a cache hider hasn't averaged 'cause some logs will say, "coords 40' off" whereas others say "coords right on". (The later when they randomly hit the same number as the hider.)


When hiders average nearly everyone says, "coords within 20'", since each finder will be there under different satellite conditions w/a different GPS.


But to answer your query Keith (knew I'd get around to it right?), the Garmins I know average would be the 60 series and 76's. (But please don't buy a GPS based on averaging capabilities alone--the limited use doesn't warrant it.)





Link to comment

blindleader Posted on Mar 19 2005, 06:49 PM

Magellans all support waypoint averaging.

All my research (many posts on this forum and google the web) indicate that not only does Magellan "support" averaging, it forces averaging on you with their "Auto Averaging" feature. At least while you are standing still.


This is not the first post where blindleader has indicated an aversion to averaging. It's possible he is right, but wouldn't it be ironic if he was using a Magellan.


I like the Garmin implementation of averaging. Select it if you want to experiment with it. Certainly the best advice to establish Cache coordinates is test them just like all subsequent searchers will. Program the candidate coordinate into your GPSr. Approach from several directions. See how close you get. Adjust as necessary. Try it on more than one day, and at different times of day.


It looks like all the Garmin "MAP" 60 and 76 units offer Waypoint Averaging. If you ever need to know about any specific GPSr unit, download the Manual.

Link to comment

I agree with Blindleader, and RJFerret as well. Averaging is not going to help you, unless you average over very long periods of time. Single points will do you better. However, as RJFerret says, if the constellation of sats is bad, averaging may help some. Usually, with SA off, averaging may only introduce more error than not. Just no way to tell. Here is what I do. I take several single point marks, watching the EPE or Accuracy Reading and marking when the number is lowest. Mind you the accuracy reading is only an estimate of how your unit is doing. It is NOT gospel. I also include an averaged mark, also waiting for the accuracy number to drop to its lowest. I then switch to "Nearest Waypoints" and see which ones come in closest. I also take a look at all the marks to see what the coordinates are and how many match or are close to each other. I then get real scientific-like and pick the one I feel best about. I usually do pretty good.



Edited by EraSeek
Link to comment

Eraseek's method is as good as you can get without averaging over very long periods of time. The emphasis is mine and can't be overstated. But just averaging over a very long time isn't going to be enough. First you need to learn about probability and statistics. It's a subject very relevant to measurement.


Answering CR. I do use a Magellan. I always defeat autoaveraging. Ironic, isn't it? Oops, just can't keep my hand away from that sarcasm button.

Link to comment
OK, so I should pretty much just keep doing what I'm doing. Try to get a day with good EPE and take a couple of readings and then manually average them.


Sound fair enough.

Every time I have found a cache in which the hider has stated they manually averaged the cache location after taking sever reading those caches have been the furthest from from the location my GPS has teken me to.

Link to comment

blindleader Posted on Mar 20 2005, 12:50 PM

I do use a Magellan. I always defeat autoaveraging.


OK Serious question here. No sarcasm. I have never owned a Magellan, but I would not rule it out. With my Garmin, I wonder why I have to collect Location data Vertical but the Electronic Compass must be Horizontal (solved that with an external Antenna). Magellan even seems to have a better implementation of the Electronic Compass.


But this is the first time I have heard about "defeating" Auotaveraging on a Magellan. Is this a commonly known "workaround" amoung Maggie fans? Is it a simple or a high tech procedure?

Link to comment
But this is the first time I have heard about "defeating" Auotaveraging on a Magellan. Is this a commonly known "workaround" amoung Maggie fans? Is it a simple or a high tech procedure?

Hi-Tech is in the eye of the beholder. Just move the unit off station a meter or two and then return it to where you're taking the fix. The few seconds of autoaveraging that takes place while your taking the fix won't make any difference.

Link to comment
With my Garmin, I wonder why I have to collect Location data Vertical but the Electronic Compass must be Horizontal (solved that with an external Antenna). Magellan even seems to have a better implementation of the Electronic Compass.

Garmin units use a two-axis magnetoresistive sensor chip. Magellan either uses a three-axis chip or a combination of two-axis and one-axis. There is also additional circuitry to compensate for tilt that provides roll and pitch information to the a/d converter. This is a bit more costly and requires more power.

Link to comment

I remember this very good thread about waypoint averaging, there are surely many others. There I explained how to obtain waypoint averaging with an eTrex legend or Vista, if you are interested in the experiment. But there was also a lot of very interesting discussion about the old question if averaging helps or not. In particular there were some very interesting links posted by Kerry, which I see now are no longer active, unfortunately. Anyway the point of that discussion was: GPS position errors on short (i.e. minutes, tens of minutes) timescales are not random. Each position reading (tipically one per second) in consumer GPS is related to the previous ones, at least because the software uses an algorhytm known as Kalman filtering exactly to avoid the measure to jump around randomly. Shortly, this filtering forces the next measurement to keep some degree of coherence with the previous ones (you may find plenty of mathematical details on net). This is generally a desirable thing, but the drawback is that the measurement error, the "wandering around" of the position, is not random, but largely systematic. As you may know, systematic errors are NOT reduced by averaging, which is effective only with symmetrical error distributions. In other terms, the error distribution on short timescales is NOT bell-shaped at all. Averaging on different days may work better, but is very time consuming, and the advantage is very small unless you have a crappy sky view in the location. Besides that, if you average together measurements taken during good-EPE and bad-EPE periods, you will most likely do worse than by taking a single measure in good EPE condition. So that the best advice is most likely to use one of the many softwares (like this one by Trimble) available for free to choose a good moment to go there and take your measurement.




Acaro of Team TAR

Link to comment

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

  • Create New...