Jump to content

Averageing


Upham

Recommended Posts

Well to be perfectly honest, I haven't actually used averages on my caches. The fun is finding the general area then hunting for the treasure. If you want to average it, just take three or four different readings from the site and see if they differ.

 

phantom icon_eek.gif

 

"Mmmmmmm,donuts" -- Homer Simpson

Link to comment

Well to be perfectly honest, I haven't actually used averages on my caches. The fun is finding the general area then hunting for the treasure. If you want to average it, just take three or four different readings from the site and see if they differ.

 

phantom icon_eek.gif

 

"Mmmmmmm,donuts" -- Homer Simpson

Link to comment

Check out this link. It might help out a little.

 

Kerry will argue that you don't really need to average if you're doing it all on one day. I know I feel better if I do.

 

If you are really concerned about giving as accurate of coordinates as you can, you might want to build in time to go out to the location on multiple days.

 

Markwell

Non omnes vagi perditi sunt

Link to comment

When I need to mark a waypoint that can be relied on I always average my position for at least 15 to 20 minutes, sometimes even more. If I am in a wooded area or in a narrow, steep-sided valley or ravine, there is always the chance that multipath errors (caused by signal reflections from the hard objects around you) will skew your position fixes. As the satellites you are receiving move across the sky, these signal reflections will also change. By averaging your position for an extended period of time you are more likely to improve the accuracy by "averaging out" these errors.

 

As Markwell stated, there are differing opinions on just about anything to do with GPS and accuracy. I tend to form a habit of those methods that work best for me. I'm currently using a Garmin GPS-12XL and a GPSmap76.

 

Cheers ...

 

~Rich in NEPA~

 

1132_1200.jpg

 

=== A man with a GPS receiver knows where he is; a man with two GPS receivers is never sure. ===

Link to comment

100 % agree with both Markwell and the Rich-n-ator ... I cache hunt usually with my Vista ... but establish caches with my 12XL or my 12Map ... so I can average it while setting up the cache ... usually around 20 minutes. Like Rich said ... it really makes a difference in a heavily tree'd area. Of course the cacher looking for your cache will not be averaging as he searches ... but I feel better establishing the cache as accurately as I feel I can.

 

348_1002.gif

Link to comment

What I've done the last two times is to place a magnetic compass at the location in question pointing to magnetic north. Turn the compass by 15 degrees toward the east and that is true direction.

 

Set a waypoint at the location using whatever accuracy you can get out of the GPSR. Then walk 20 and 40 feet in all four directions collecting waypoints at each consistant range. The two east, center, and two west waypoints should have very close longitude and the two north, center and two south waypoints should have very close latitude.

 

I mark the waypoints as C, N1, N2, W1, W2, ... etc. I also plug the accuracy reading into the altitude field of the waypoint.

 

When I get home, I export the waypoint information to a CSV and import that into an Excel spreadsheet that I wrote to do the averaging.

 

Can't be done on some terrain and it is a lot of extra work. If there is an easier way to average on the Vista, I'd love to hear about it.

Link to comment

quote:
Originally posted by rut:

If you think I should average can you advise me on how to do it?

Thanks

Bob


To average with the Map76, you hold down the [enter] key to set up the mark...then hit the [menu] key, and you will get an option to average. You select that and it will start a count. You can hit the SAVE at any time, or wait for the count to slow or stop (it's found it's best average). I usually average, but just for a minute or less to make sure I've eliminated the possiblity of any wild signal, unless I am having a difficult time picking up a good signal.

 

2573_200.jpg

Link to comment

The difference can be about 3/5th's of 5/8th of .... but if it makes one "feel" better then certainly do it.

 

Currently trying to put together a different angle on averaging and it would appear that after a certain period of time (and fairly short at that) the accuarcy appears to/can go against the purpose. The thing is one doesn't know if it is or not which can make it just a "feel good" excerise.

 

Cheers, Kerry.

 

I never get lost icon_smile.gif everybody keeps telling me where to go icon_wink.gif

 

[This message was edited by Kerry on March 09, 2002 at 10:45 PM.]

Link to comment

The difference can be about 3/5th's of 5/8th of .... but if it makes one "feel" better then certainly do it.

 

Currently trying to put together a different angle on averaging and it would appear that after a certain period of time (and fairly short at that) the accuarcy appears to/can go against the purpose. The thing is one doesn't know if it is or not which can make it just a "feel good" excerise.

 

Cheers, Kerry.

 

I never get lost icon_smile.gif everybody keeps telling me where to go icon_wink.gif

 

[This message was edited by Kerry on March 09, 2002 at 10:45 PM.]

Link to comment

Just wondering, doesn't the averaging function induce more errors ever since Clinton turned off Selective Availabity?

 

The averaging function on GPSR's, to my understanding, are to average out the errors caused by SA.

 

With the SA turned off, then I believe it is making a mockery of the averaging algorithm.

 

I tried marking a waypoint with and without the averaging function after SA was turned off 2 years ago, and was surprised to discover that the average one was about 10m off my waypoint as compared to the non-averaged one.

 

Any comments/thoughts?

Link to comment

Averaging around trees can be problematic. It would be best to do a few 5 minute averaged waypoints, at different times, then average those waypoints. I found that averaging for 5 minutes on my Garmin GPS V, gives me a good reading, for when i place a cache. Problem with trees and averaging, is that the GPS unit may switch satellites during an averaging, caused by intermittant blocking of the Sats by the trees.

 

I have a cache that I averaged a waypoint over a Survey Marker, and it was within 6 ft of it(GPS V), and 16 feet of the marker, with the GPS 315. The Survey Marker was out in the open, so my GPS V gave me a good reading, when i averaged for 5 minutes.

 

Here is the cache:

NGS Markers & GPS Accuracy Check Geocache

 

5_Rubik.gif

Link to comment

quote:
Originally posted by sporthunter:

EraSeek I hear you say that with your map76 it will stop when it gets it's most accrete reading, With my emap it just keep's counting butt I've maybe waited five minutes ten, If I was to let it go would it stop also at the best average?

 

sporthunter


 

With my Garmin 12 you can watch the accuracy reading it provides jump around and eventually stablized in a lower reading when it has averaged sufficiantly. My Map76 counts, and I have noticed that sometimes it will stop after an amount of time. I assume it is doing the same thing as how it stops counting quicker with good coverage. Moun10Bike has not noticed this on his. Maybe it is just a bug. I plan to play with it a bit more while paying attention to this feature to confirm that that is what it is actually doing.

 

2573_200.jpg

Link to comment

quote:
Originally posted by durianwool:

Just wondering, doesn't the averaging function induce more errors ever since Clinton turned off Selective Availabity?

 

The averaging function on GPSR's, to my understanding, are to average out the errors caused by SA.

 

With the SA turned off, then I believe it is making a mockery of the averaging algorithm.

 

I tried marking a waypoint with and without the averaging function after SA was turned off 2 years ago, and was surprised to discover that the average one was about 10m off my waypoint as compared to the non-averaged one.

 

Any comments/thoughts?


 

Yes I think averaging was more in tune when SA was on but since then those units with upgradeable software probably have had changes made in recent releases (if one as upgraded the software of course?). Some may still be running pre SA and some post SA software?

 

Averaging can certainly make the position worse (can also make it better icon_smile.gif) but the real problem is one doesn't really know.

 

Cheers, Kerry.

 

I never get lost icon_smile.gif everybody keeps telling me where to go icon_wink.gif

Link to comment

quote:
Originally posted by durianwool:

Just wondering, doesn't the averaging function induce more errors ever since Clinton turned off Selective Availabity?

 

The averaging function on GPSR's, to my understanding, are to average out the errors caused by SA.

 

With the SA turned off, then I believe it is making a mockery of the averaging algorithm.

 

I tried marking a waypoint with and without the averaging function after SA was turned off 2 years ago, and was surprised to discover that the average one was about 10m off my waypoint as compared to the non-averaged one.

 

Any comments/thoughts?


 

Yes I think averaging was more in tune when SA was on but since then those units with upgradeable software probably have had changes made in recent releases (if one as upgraded the software of course?). Some may still be running pre SA and some post SA software?

 

Averaging can certainly make the position worse (can also make it better icon_smile.gif) but the real problem is one doesn't really know.

 

Cheers, Kerry.

 

I never get lost icon_smile.gif everybody keeps telling me where to go icon_wink.gif

Link to comment

Thanks EraSeek, I didn't know I had that function. I tried it today and I just watched it averaging when I thought it was good I stopped it. I think it worked great. I will go back soon to check. I placed my first cache but when I was leaving I found a cave. It's not real big or deep but it's real cool. All the way home I was arguing with myself because I like both spots I finally thought to make it a multiple cache. So I have to go back I'll check it then.

Thanks again,

Bob

Link to comment

One way to do it, is if some people had some money to rent a Survey Grade GPS, to compare with a few consumer model GPS units, in the open and in the woods.

 

Is this an idea?

This is one way to see if there are any consumer GPS units that can do Averaging well, or not.

 

One way to test the Averaging function on many GPS units is like in this Example:

http://www.geocaching.com/seek/cache_details.asp?ID=10951

This Survey Marker is a good way to test many different GPS units. There are several log on this cache page.

 

5_Rubik.gif

Link to comment

quote:
Originally posted by rut:

If you think I should average can you advise me on how to do it?


 

This is how I've been averaging with my Venture lately.

I mark the spot I've selected (Big surprise) Naming it whatever-the-heck.

The I click "OK"

I re-select the waypoint by name and look at the distance field.

There should be something there, 20ft or whatever the difference is between where you are and where it thinks you are.

Use clik-stick to select the secondary menu and click on the sixth menu item "Reposition Here"

Set the unit down and wait a few minutes until the distance shows something and press the clik-stick again, which zero's the distance number again.

Keep repeating until things are stable.

Try using more than one waypoint name for added security of accuracy.

If you have the time and desire to be sure, you can return at a later time to verify the coordinates.

No matter how many times you do so, a certain amount of inaccuracy will remain.

Link to comment

quote:
Originally posted by durianwool:

The averaging function on GPSR's, to my understanding, are to average out the errors caused by SA.


 

Unlikely. SA was designed specifically to defeat averaging. When SA was on, a track log of a stationary object would wander all of the place, but slowly (about 2-3 mph) and with some jumps as your receiver switched to a different PRN. A fifteen or twenty minute average of such a progression could be off from the actual location by as much as or more than any of the individual fixes that went into it. The only

way averaging was any good under SA was if you averaged several hours worth of fixes, to give the error plenty of time to wander all over the target area. Even then, visual averaging of the resulting path was often more accurate than mechanical.

Link to comment

quote:
Originally posted by Geoffrey:

One way to do it, is if some people had some money to rent a Survey Grade GPS, to compare with a few consumer model GPS units, in the open and in the woods.

 

Is this an idea?

This is one way to see if there are any consumer GPS units that can do Averaging well, or not.

 


 

No won't really prove much as a survey grade GPS by itself won't really do all that much better.

 

Cheers, Kerry.

 

I never get lost icon_smile.gif everybody keeps telling me where to go icon_wink.gif

Link to comment

quote:
Originally posted by Geoffrey:

One way to do it, is if some people had some money to rent a Survey Grade GPS, to compare with a few consumer model GPS units, in the open and in the woods.

 

Is this an idea?

This is one way to see if there are any consumer GPS units that can do Averaging well, or not.

 


 

No won't really prove much as a survey grade GPS by itself won't really do all that much better.

 

Cheers, Kerry.

 

I never get lost icon_smile.gif everybody keeps telling me where to go icon_wink.gif

Link to comment

quote:
Originally posted by Warm Fuzzies - Fuzzy:

 

Unlikely. SA was designed specifically to defeat averaging. ... The only way averaging was any good under SA was if you averaged several hours worth of fixes, to give the error plenty of time to wander all over the target area.


 

A mathematician named Wilson did quite a bit of research on averaging under SA and his work showed that averaging under SA would improve accuracy if the averaging period was longer than about 6 minutes. The longer the period of averaging, the better the results, but even just 30 minutes of averaging would reduce error by more than 50%. It may seem counter-intuitive, but Wilson has shown that with SA off, averaging is much less effective in improving relative accuracy.

 

FWIW,

CharlieP

 

[This message was edited by CharlieP on March 14, 2002 at 09:20 PM.]

Link to comment

quote:
Originally posted by CharlieP:

A mathematician named Wilson did quite a bit of research on averaging under SA and his work showed that averaging under SA would improve accuracy if the averaging period was longer than about 6 minutes. The longer the period of averaging, the better the results, but even just 30 minutes of averaging would reduce error by more than 50%.


Sure, but that's just reducing the error from 100 meters to 50 meters. It's still nothing like what I'd call a good fix.

Link to comment

quote:
Originally posted by CapnGadget:

I have a basic yellow etrex...usually I walk about 50 - 100 yards from the cash, refix and walk back to the cache and take note of the co-ordinates...I do this 4 times to the cardinal points and then calculate the average and this is what I post. icon_cool.gif


 

I don't think the system actually changes enough over time these days (without SA) for that to be entirely effective.

 

Generally the relative accuracy between those cardinal points will be close but the overall absolute accuracy of all those points (to the intended location) will still be an unknown.

 

The following is a practical real world example similar to what you describe and what some other users have previously nicknamed "golf ball averaging"

 

icon_smile.gif everybody keeps telling me where to go icon_wink.gif

Link to comment

quote:
Originally posted by CapnGadget:

I have a basic yellow etrex...usually I walk about 50 - 100 yards from the cash, refix and walk back to the cache and take note of the co-ordinates...I do this 4 times to the cardinal points and then calculate the average and this is what I post. icon_cool.gif


 

I don't think the system actually changes enough over time these days (without SA) for that to be entirely effective.

 

Generally the relative accuracy between those cardinal points will be close but the overall absolute accuracy of all those points (to the intended location) will still be an unknown.

 

The following is a practical real world example similar to what you describe and what some other users have previously nicknamed "golf ball averaging"

 

icon_smile.gif everybody keeps telling me where to go icon_wink.gif

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.

Guest
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.

Loading...
×
×
  • Create New...