Jump to content

Using benchmarks to test consumer GPS accuracy


Mapping Support

Recommended Posts

I have searched through this forum and elsewhere online and come up dry.

Surely someone is compiling this info somewhere but I am having a tough time finding it!

 

Here's the concept.

 

Stand right at a benchmark for ~1 minute.

You already have a GOTO set, right?

 

Now read off the distance to your waypoint.

That is a measure of the inaccuracy of your GPS for this specific test.

 

Of course if you repeat the test at a different time then you will likely get a different result since the satellites will be in a different 'constellation' and atmospheric conditions will have changed.

 

I think this would be a fascinating way to compare the accuracy of any GPS-enabled device.

 

You can even shave the odds by testing when the satellites are in a favorable position:

http://satpredictor.navcomtech.com

 

I got interested in this accuracy issue since I am building a system to help people find survey stakes.

If anyone knows of a site collecting this info, please point me to it.

Link to comment

Go the NGS web site www.ngs.noaa.gov and select "Survey Mark Datasheets" from the left side of the page, then "DATASHEETS" from the next page. This takes you to the NGS DATASHEET RETRIEVAL PAGE. Once you've determined how you want to retrieve the data on the following page select under Date Type Desired select GPS Sites Only. This will provide you with a listing of the passive control stations with the highest horizontal positional accuracy - in most cases better than 2-3 cm. Also by selecting this option you will be assured of getting a listing of stations that were at least once occupied with geodetic quality GPS and should (I say should) have the fewest obstructions.

Link to comment

You have to understand though, that most benchmarks were placed and their position recorded decades before man even put a satellite into orbit let alone developed a GPS system. The key is to look and see if the Datasheet page lists the coordinates as SCALED or ADJUSTED as described below. As long as you are reading from a Datasheet with ADJUSTED coordinates you might be OK.

 

1. If it is scaled coordinates, the numbers can be off by more than 600 feet. What a scaled coordinates is, is where a person took a topographic map, with a ruler, and made the best estimation they could to get the coordinates. You need to remember that most benchmarks were set out long before there was anything such as global positioning satellites (GPS).

 

When entering the given coordinates in your gpsr for a scaled benchmark, then it will get you to within about 600 feet that way. You will need to use the description given on the page (like distances from roads, houses, witness posts, and other features) to find the benchmark. (Would you want to test your accuracy this way?)

 

2. Now if it has adjusted coordinates, then the numbers are more accurate than you can get with a civilian grade handheld gps unit. Precision equipment was used to arrive at the coordinates for this. You can usually get to within about 20 feet using your handheld. Your only off, because your gpsr is not as accurate as needed. A benchmark with adjusted coordinates will greatly narrow down your search area, and give you a pretty close place to start.

Link to comment

Yes CoolCow, I intend to only use benchmarks that are adjusted.

I am going to enter the benchmark coordinates into my handheld and hit GOTO.

When I find the BM then I am going to stand right there for a minute or so and then look at the distance for the GOTO.

That distance will be the amount of error for that test.

 

I posted since I was curious if anyone knows of a site or thread somewhere that has results from people who have done this test.

Link to comment

I've done that hundreds of times.

Usually the GPSr is a few feet off.

 

I have done that with my GPSr maybe 30 times. With WAAS on and also recieving the GLONASS satellites, I have found the standard deviation of the error distance to be around 6 feet. This would put the 90% confidence distance around 10 to 12 feet, if my memory serves me. This is near the manufacturer's stated accuracy.

Link to comment

My results with a Garmin 76s (about 2000 vintage unit) under good conditions (unobscured sky and good satellite constellation) are similar to Birder's. I have found that I sometimes have to stand there several minutes if the elevation is still slowly settling in before the unit gives me its best indicated accuracy numbers.

 

I once did some experiments with long-term averages, and found that occasionally even an hour-long average was 20 feet off. Must be finer-grain variations that WAAS doesn't reflect.

 

By averaging track log data for hours on multiple days, I was able to get errors near the computational resolution of that unit, a rounding or truncation somewhere around 2 or 3 feet, deep in its mathematical engines and not just in the display.

 

At that time, all recreational grade units treated NAD83 and WGS84 as equal (the null transform from an ancient military document of the early GPS days). The results are then, of course, WGS84 no matter which datum is set. I don't know if this situation remains true for current models. The datum conversion is 2 to 3 feet difference in the US (see NGS program HTDP), which is negligible unless you are fussing with long averages.

Link to comment

On one of my last outside test, I compared a SporTrak Pro, and a Meridian Color, side by side with a ProMark 3, all setup to display in centimeters, in the User Grid. They were left sitting on the top of a wooden storage unit for about 20 minutes, averaging, before I checked on them. During the next 10 min the SporTrak was at worst, 32 cm's EW, and 28 cm's NS. At best it showed 3 cm's EW, and 2 cm's NS difference from the PM 3. The Meridian was close to the same, being 37 cm's EW and 32 cm's NS, at worst, and 5 cm's EW and 3 cm's

NS at best. All three times that I have gone over to a nearby survey station, the PM 3 was within 1.5 feet of the point L/L. I always felt that under ideal conditions, the older Maggies were capable of close to a meter in accuracy, and sometimes sub meter.

When I set the Garmin GPS 60 (yellow one) and the PM 3 to hddd.ddddd. they both read the same numbers, and the 60CSx Sirf read one number off from the PM 3. Garmin makes it "tuff" trying to do the small stuff in the User Grid, even though I can cheat in the Scale and get a 60 to show in feet. Then there is that thing of the Lat being tie to the Equator, so one has to play with minus numbers. The 62s, that I had for a short while, was no better. I'm still trying to figure out how to setup the PM 3 to be able to compare it with the 60 series Garmins in Grid Feet.

When I let a 60CSx, connected to the external antenna on the roof, average in Grid Feet for three nights, it showed the same for two nights, and was off one foot the last night.

I've done a lot of other testing in the backyard with, and without, the units connected to an ext ant, ground plane, pole, tripod setup. Testing is endless, and I'm still trying and learning.

Thanks for the HDTP site, as mine got lost some time ago in a crash. In recent times I've been using the first MobileMapper, as it will display the "real" NAD 83. At my location, WGS 84 is about 4 ft W of 83, and 2 ft N of 83.

Link to comment

I might try this next time I'm in Menlo Park, which should be next week. USGS has a survey mark in their flagpole circle, and they hand out printed sheets giving its coordinates. Here's some info I posted about it some years ago, in case anyone reading this thread is in the vicinity and wants to calibrate their GPS receivers with that mark:

 

Info on WMC 1994

 

And here are some photos I took of it and the surrounding area:

 

WMC 1994

Link to comment

Just set up a cache on a benchmark to give users the 'false' opportunity to check their GPS "accuracy" :)

Accuracy of a GPS is in itself grossly inaccurate ! It varies and is not repeatable. It largely depends on the satellite configuration among other things. At time, your GPS will be pin point accurate, at other 10m off. That doesn't mean, it became bad !

 

But using a benchmark for a cache at least takes your GPS accuracy out of the game, provided the benchmark coord. have been properly updated.

Link to comment

Just set up a cache on a benchmark to give users the 'false' opportunity to check their GPS "accuracy" :)

Accuracy of a GPS is in itself grossly inaccurate ! It varies and is not repeatable. It largely depends on the satellite configuration among other things. At time, your GPS will be pin point accurate, at other 10m off. That doesn't mean, it became bad !

 

But using a benchmark for a cache at least takes your GPS accuracy out of the game, provided the benchmark coord. have been properly updated.

 

GeoCachers,

 

This might be a good time again clarify terminology and accuracy of Geodetic survey points.

 

Copy/Paste excerpts from the 'Introduction' on the Benchmark Hunting home page - my notations bold:

 

Vertical Control Marks

These are the true "bench marks". Generally the words BENCH MARK will be printed on them near their rim if the mark is the disk type. Many vertical control marks are not the disk type, however, and can include bolts, rivets, chiseled squares, chiseled crosses, etc.

 

Most are published to whole second accuracy only. DDDMMSS (about a hundred feet)

 

Horizontal Control Marks

There are several types of horizontal control marks.>>>>>

 

Triangulation Stations (a Type of Horizontal Control)

A triangulation station is usually a metal disk with a small triangle in its center. Generally the words TRIANGULATION STATION will be printed on them near their rim. Its position has been determined by measuring distances and angles from other stations. Triangulation stations, also called tri-stations, are typically associated with nearby reference mark disks and an azimuth mark disk. More about that below.

 

Should be published to five decimal second accuracy.DDDMMSS.sssss (under an inch with survey grade units).

As Suscrofa pointed out all GPSr accuracy is dependent on satellite geometry. Users of recreational units are kidding themselves if they think that they can get repeatable readings under several feet.

 

Much more good information in the 'Pinned' topics at the top of this forum.

 

kayakbird

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...