Factors affecting accuracy.

[+tec_64]

I've been checking the consistancy of readings of my GPS by repeat visits to a high-accuracy benchmark.

 

For the most part, I've been getting very consistant repeat readings. I use the same procedure and setup each time (eg. GPS on for at least 20 minutes for almanac update to load, clear sky, etc. etc .). I use Trimble Planning software with the latest Keps to determine the best HDOP (usually less that 0.8) times.

 

Again, for the most part, my readings repeat themselves within a metre. Even with all that, every once in a while I get a coord that is out of whack. Any ideas? Aliens? I didn't see any black helicopters. I had my foil hat on.

 

I'm not anal about it. It was just a curiousity and I had some (too much?) time on my hands.

[+Night Stalker]

I have been told that the time of day can make a difference because the satelites may be situated in different parts of the sky and this can effect accuracy.

[+GerIRL]

My guess would be time of day and actual satellite configuration at the exact time you're checking in. We use Leica survey grade GPS equipment at work. Usually, our RTK receivers (Real Time Kinetic) are good to 1.5cm in both position and elevation when processed using 1 second data collected at a known base station (NJIT in Newark, NJ). Sometimes, we simply can't run RTK because there are a couple of satellites down, or the configuration is simply not suitable. For the most part, mornings seem to be better than afternoons. I imagine the accuracy of your consumer handheld is even more susceptible to these variations.

[+KD7MXI]

I've been checking the consistancy of readings of my GPS by repeat visits to a high-accuracy benchmark.

 

For the most part, I've been getting very consistant repeat readings. I use the same procedure and setup each time (eg. GPS on for at least 20 minutes for almanac update to load, clear sky, etc. etc .). I use Trimble Planning software with the latest Keps to determine the best HDOP (usually less that 0.8) times.

 

Again, for the most part, my readings repeat themselves within a metre. Even with all that, every once in a while I get a coord that is out of whack. Any ideas? Aliens? I didn't see any black helicopters. I had my foil hat on.

 

I'm not anal about it. It was just a curiousity and I had some (too much?) time on my hands.

 

i noticed on my etrex legend the accuracy goes way down when i turn on the advanced features such as maps and waas

[+julianh]

i noticed on my etrex legend the accuracy goes way down when i turn on the advanced features such as maps and waas

Note that the EPE reported by Garmin units is only an estimate of probable accuracy, not a calculation of actual accuracy. Don't pay too much attention to the reported figure, other than that a high EPE means your unit estimates its current accuracy is quite low, and a low EPE figure means your unit believes its current accuracy is quite good.

 

Generally, turning WAAS on would be expected to reduce the EPE figure; that is, WAAS should in theory increase accuracy, not decrease it. If you are seeing a decrease in accuracy, there are two possible reasons I can think of (there may be others too!)

 

1. Dumb luck - you happened to turn WAAS on at a time of day when the satellites happened to be configured overhead so as to give you lower accuracy, and / or you happened to try using WAAS when your location meant you couldn't get a high accuracy position fix (e.g. multi-path errors, overhead shielding due to trees, etc).

 

2. In addition to its primary user function of correcting for ionosphere errors, etc, WAAS provides information about satellite integrity etc which is not available without WAAS. In theory, this means your GPSr gets a more reliable fix, including a better estimate of the actual likely accuracy. If your unit reports higher EPE when WAAS is turned on, this may in fact be more a reliable estimate of current error than the "optimistic" estimate of EPE that your unit makes without the benefit of the WAAS corrections.

 

Turning maps on and off should not affect the EPE (the unit's estimate of current accuracy). However, the radius of the error circle shown on the map screen can go up and down depending on which map set you are viewing. This is because the size of the error circle depends on BOTH the EPE (the unit's estimate of its position fix accuracy) AND the precision of the map. Highly detailed maps will often show a smaller error circle than low res maps (such as the base map, or 100,000 scale topo maps etc), because the error circle tries to give you a display of the confidence of your position AS SHOWN ON THE CURRENT MAP. With high res maps loaded, the built-in map error is very small, so the radius of the error circle is closely related to the current EPE (estimate of GPS location accuracy). With low-res maps loaded, the diameter of the error circle increases, because map features themselves could be out of position by several hundred metres, so the error circle has to take into account the built-in map errors as well as the EPE.

 

Hope this helps!

[+KD7MXI]

i noticed on my etrex legend the accuracy goes way down when i turn on the advanced features such as maps and waas

Note that the EPE reported by Garmin units is only an estimate of probable accuracy, not a calculation of actual accuracy. Don't pay too much attention to the reported figure, other than that a high EPE means your unit estimates its current accuracy is quite low, and a low EPE figure means your unit believes its current accuracy is quite good.

 

Generally, turning WAAS on would be expected to reduce the EPE figure; that is, WAAS should in theory increase accuracy, not decrease it. If you are seeing a decrease in accuracy, there are two possible reasons I can think of (there may be others too!)

 

1. Dumb luck - you happened to turn WAAS on at a time of day when the satellites happened to be configured overhead so as to give you lower accuracy, and / or you happened to try using WAAS when your location meant you couldn't get a high accuracy position fix (e.g. multi-path errors, overhead shielding due to trees, etc).

 

2. In addition to its primary user function of correcting for ionosphere errors, etc, WAAS provides information about satellite integrity etc which is not available without WAAS. In theory, this means your GPSr gets a more reliable fix, including a better estimate of the actual likely accuracy. If your unit reports higher EPE when WAAS is turned on, this may in fact be more a reliable estimate of current error than the "optimistic" estimate of EPE that your unit makes without the benefit of the WAAS corrections.

 

Turning maps on and off should not affect the EPE (the unit's estimate of current accuracy). However, the radius of the error circle shown on the map screen can go up and down depending on which map set you are viewing. This is because the size of the error circle depends on BOTH the EPE (the unit's estimate of its position fix accuracy) AND the precision of the map. Highly detailed maps will often show a smaller error circle than low res maps (such as the base map, or 100,000 scale topo maps etc), because the error circle tries to give you a display of the confidence of your position AS SHOWN ON THE CURRENT MAP. With high res maps loaded, the built-in map error is very small, so the radius of the error circle is closely related to the current EPE (estimate of GPS location accuracy). With low-res maps loaded, the diameter of the error circle increases, because map features themselves could be out of position by several hundred metres, so the error circle has to take into account the built-in map errors as well as the EPE.

 

Hope this helps!

 

my map screen likes to show that im up to 2 blocks from my current location while traveling in a car when mapsource uploaded maps are turned on - when the uploaded maps are off it update my location almost emediatly - my lat long is fine but my map updates extremely slow - waas and no waas

 

when sitting perfectly still the map shows my location perfectly even when indoors - waas and no waas

 

my gps shows a D on all available birds when waas is turned on - i thought only 3 where waas - my gps shows there all waas capable - talk about strange

Edited November 13, 2008 by KD7MXI

[japa-fi]

my gps shows a D on all available birds when waas is turned on - i thought only 3 where waas - my gps shows there all waas capable - talk about strange

Those with D are "normal" GPS satellites which are corrected/adjusted by WAAS data. Ie. they are not WAAS satellites.

[+twolpert]

Since he was using a benchmark, I am pretty sure the OP was comparing coordinates on multiple visits, not EPE. Unless circumstances (such as the amount of foliage) change, differences are most likely due to differing constellations. In any case, readings repeatable to within a meter on a consumer handheld are remarkably good. Larger variations are to be expected -- at least some of the time.

[+trainlove]

Connect your GPSr to a computer.

Run HyperTerminal or other Terminal emulator software.

Turn on NMEA on your GPSr.

 

You will see the PDOP, HDOP and VDOP numbers, which are not meaningless calculations, but are actual precision factors.

 

You will notice that those change every second. Well that's why EPE changes every second since EPE is mostly just a constant times PDOP.

 

P.S. When using benchmarks, be sure not to use Scaled ones, and to use the ones that are 'high precision' position locations, mostly Adjusted ones but be sure that it's First Order or similarily high.

[+SimbaJamey]

Rain.

 

 

On rainy days I usually have serious accuracy issues with my Colorado.

Not just when I'm out caching...

I have a stump in my yard that I've been using to conduct my own tests on the Co. for a while now. I leave it on the stump for a bit and then mark a waypoint. Most days it does really well (within 10' of the others), but on rainy days I often get points that are as much as 70' away no matter how long I let it settle.

[+twolpert]

Rain.

 

 

On rainy days I usually have serious accuracy issues with my Colorado.

Not just when I'm out caching...

I have a stump in my yard that I've been using to conduct my own tests on the Co. for a while now. I leave it on the stump for a bit and then mark a waypoint. Most days it does really well (within 10' of the others), but on rainy days I often get points that are as much as 70' away no matter how long I let it settle.

Wet foliage overhead? Coating of water on the unit itself? Solid wall of water downpour? Significant multi-path conditions in the neighborhood anyhow? Almost has to be one of those things. Generally speaking, clouds, rain, snow, fog, etc. do not attenuate GPS signals enough to cause observable accuracy issues. On the other hand, "solid" water -- even a thin coating on leaves or the unit itself -- can cause much more significant attenuation.

 

I don't doubt that you actually see problems with your reported position. But I think something in addition to rainfall must be involved. My money would be on significant multi-path conditions coupled with water on the unit. But that's just an opinion, for what it's worth.

[japa-fi]

I have started recording a position of a certain pole, which is at quite open area. I have measurements on 4 different days. One of the measurements differ drastically, it's 28 meters from others.rest of the measured locations are inside a 4 meter radius circle.

today two measurements with a time difference of 50 minutes and cloudy sky turning to slight drizzle the difference was 5 meters.

[japa-fi]

[dupe]

Edited November 13, 2008 by japa-fi

[+NOV8TR]

"trainlove" said:

"You will see the PDOP, HDOP and VDOP numbers, which are not meaningless calculations, but are actual precision factors.

 

You will notice that those change every second. Well that's why EPE changes every second since EPE is mostly just a constant times PDOP."

 

Wow,All those acronyms! I just had to know,,,,, I found this old site that helped a bit but, would still like to know more!

link

Edited November 13, 2008 by 1XL-on-XR650L

[+trainlove]

One of the measurements differ drastically, it's 28 meters from others.rest of the measured locations are inside a 4 meter radius circle.

In theory, there's nothing saying that you could have one reading, or even 5% of your readings being hundreds of miles away from the 'correct' position.

Of course that's really bad, but the GPS system is only guarranteed to have something like 95% of your readings being within 10 meters of ground zero.

 

My GPS once, for nearly NO reason, thought I was in the middle of the Great Lakes instead of Massachusetts. Well that was because they turned off WAAS PRN122 and somehow that temporarily corrupted my almanac.

It was fun to see my speed slowly change from over a million miles per hour to mere hundreds of miles per hour during the time it took for a complete re-acquisition of the almanac. Yes, my GPS was able to display such a speed even though it's only spec'ed to be valid up to something like 500MPH.

[Ken in Regina]

"trainlove" said:

"You will see the PDOP, HDOP and VDOP numbers, which are not meaningless calculations, but are actual precision factors.

 

You will notice that those change every second. Well that's why EPE changes every second since EPE is mostly just a constant times PDOP."

 

Wow,All those acronyms! I just had to know,,,,, I found this old site that helped a bit but, would still like to know more!

link

Stick any of those acronyms, or just DOP (Dilution Of Precision), plus the word "wikipedia" in the Google search box and you'll get a good list with Wikipedia at the top of the list of search results.

 

...ken...

Edited November 13, 2008 by Ken in Regina

[+iamasmith]

...

my gps shows a D on all available birds when waas is turned on - i thought only 3 where waas - my gps shows there all waas capable - talk about strange

 

Not sure about the mapping symptom without seeing it more but your observation about the WAAS (or EGNOS, both behave the same) is explained as follows.

 

The three satellites that you talk about have a special purpose. Basically ground monitoring stations tightly monitor their position against known fixed locations and also monitor the data being generated by the other GPS satellites.

 

Correctional information is generated for ALL the satellites and sent to these three from where it is sent to your WAAS capable receiver. The WAAS capable receiver then applies the corrections to ephemeris data that the GPS satellite sends that describes its path and location. That's why for 3 WAAS satellites you get correction on most if not all of them.

 

Occasionally you see ones that don't have a D when some do. I don't have enough information about why this is right now but I can only assume that this is because the WAAS network has deemed the current ephemeris to be accurate and is not bothering sending corrections

 

BTW Almanac is just used to do initial location based upon where it believes the satellites last were and to prep the unit for receiving expected satellites. AFAIK all actual locations are based upon ephemeris.

 

Waiting 20 mins is more a factor for gaining an accurate time signature since GPStime doesn't include 'leap seconds' - it's just a count of number of seconds since the launch of the service. Currently there are 22 leap seconds since that epoch (IIRC) that need to be accounted for. Every 15 mins the correction information is sent out. Once the unit has received that the clock is pretty accurate indeed but up to that time it may be 22 seconds out.

 

Ephemeris data uses GPStime and the the timing is purely for time application not navigational so waiting the 20 mins for time sync shouldn't effect navigation AFAIK. I'm still getting all the background though so if anybody has more detailed info I just love that kind of stuff

 

Regards,

 

-Andy

Edited November 13, 2008 by iamasmith

[+Team CowboyPapa]

Do a search on Klemmer. He is extremely knowlegeable in this subject and has posted herein previously.

[+KD7MXI]

...

my gps shows a D on all available birds when waas is turned on - i thought only 3 where waas - my gps shows there all waas capable - talk about strange

 

Not sure about the mapping symptom without seeing it more but your observation about the WAAS (or EGNOS, both behave the same) is explained as follows.

 

The three satellites that you talk about have a special purpose. Basically ground monitoring stations tightly monitor their position against known fixed locations and also monitor the data being generated by the other GPS satellites.

 

Correctional information is generated for ALL the satellites and sent to these three from where it is sent to your WAAS capable receiver. The WAAS capable receiver then applies the corrections to ephemeris data that the GPS satellite sends that describes its path and location. That's why for 3 WAAS satellites you get correction on most if not all of them.

 

Occasionally you see ones that don't have a D when some do. I don't have enough information about why this is right now but I can only assume that this is because the WAAS network has deemed the current ephemeris to be accurate and is not bothering sending corrections

 

BTW Almanac is just used to do initial location based upon where it believes the satellites last were and to prep the unit for receiving expected satellites. AFAIK all actual locations are based upon ephemeris.

 

Waiting 20 mins is more a factor for gaining an accurate time signature since GPStime doesn't include 'leap seconds' - it's just a count of number of seconds since the launch of the service. Currently there are 22 leap seconds since that epoch (IIRC) that need to be accounted for. Every 15 mins the correction information is sent out. Once the unit has received that the clock is pretty accurate indeed but up to that time it may be 22 seconds out.

 

Ephemeris data uses GPStime and the the timing is purely for time application not navigational so waiting the 20 mins for time sync shouldn't effect navigation AFAIK. I'm still getting all the background though so if anybody has more detailed info I just love that kind of stuff

 

Regards,

 

-Andy

 

today at a cache location my gps showed i was 30 feet from ground zero with waas on at ground zero - it showed i was 3 to 6 feet from ground zero with waas off at ground zero

 

when i go caching without waas i find caches like crazy - when waas is on when caching i get huge number of no finds

Edited November 14, 2008 by KD7MXI

[Ken in Regina]

today at a cache location my gps showed i was 30 feet from ground zero with waas on at ground zero - it showed i was 3 to 6 feet from ground zero with waas off at ground zero

 

when i go caching without waas i find caches like crazy - when waas is on when caching i get huge number of no finds

The most likely explanation for that is that whoever placed those caches did not have WAAS enabled and when you have WAAS turned off you are seeing approximately the same error as they had.

 

Sometimes you can be too accurate.

 

Perhaps cache owners should be encouraged to note whether the location was recorded with or without WAAS.

 

...ken...

[+KD7MXI]

today at a cache location my gps showed i was 30 feet from ground zero with waas on at ground zero - it showed i was 3 to 6 feet from ground zero with waas off at ground zero

 

when i go caching without waas i find caches like crazy - when waas is on when caching i get huge number of no finds

The most likely explanation for that is that whoever placed those caches did not have WAAS enabled and when you have WAAS turned off you are seeing approximately the same error as they had.

 

Sometimes you can be too accurate.

 

Perhaps cache owners should be encouraged to note whether the location was recorded with or without WAAS.

 

...ken...

 

true

[+tec_64]

Here's the details of my "experiment", in case anyone cares.

 

http://www.personainternet.com/mile77/3709562.htm

 

tec_64

 

Comments welcome.

[+NOV8TR]

Here's the details of my "experiment", in case anyone cares.

 

http://www.personainternet.com/mile77/3709562.htm

 

tec_64

 

Comments welcome.

 

Nice job!

I have been considering just such a test, but without taking into consideration all the "optimum signal time/day" constraints. I have an "adjusted horizontal" bench mark within 1 1/5 miles of my home. My plan was to visit the site as often as my marriage/work situation permitted (most often Saturday mornings between 8:00am and 11:00am). The thought was to record the difference between the EPE of my Garmin Vista HCx and the actual (documented known) position of the bench mark. My true interest was to put some logic to the EPE reading on my GPS rather than just "taking their word for it". If I could record that delta between reported EPE and actual error for maybe 10~15 data points, then I might get get a fairly good feel for what reported EPE really meant in the real world (outside of the Garmin software lab).

Anyway,,,, it is nice to see there is someone else out there that is curious enough to "do the work", and record the results!

Again,, nice job!

[+mchaos]

Lots of things can effect a reading. Trees, their leaves can bounce signals. Buildings can do the same. Clouds can also have a minimal effect on gps.

 

You may also take into consideration, the GPS unit. Software on it, and how powerful of a receiver and antenna it has.

 

I am sure if you had a first generation Military unit, It would probably work much better and have complicated algorithms to counter these types of interferences.

 

I came to some simple conclusions when I would get some spearatic readings out of no where, and notice the wind was blowing pretty hard and I was near trees.