8

I've already read this and this to get some background knowledge on GPS tools, but I'm still left with some questions:

  1. Is there any way I can get a device that consistently provides accuracy to within 10 feet in less than a minute or two?

  2. I've messed around with several iPhone apps, but haven't had very good luck thus far. What's the most accurate a phone can be?

  3. If I can't have accuracy to less than 10 feet within two minutes, what's the closest I can get to that goal? (Both in terms of time and accuracy)

I only need to be able to accurately determine the coordinates of specific locations, so I don't need any additional features--even poor battery life isn't a big deal.

8
  • 3
    Averaging data points will help to some extent. I've done this with a dedicated GPS (normal pocket Garmin) and a laptop. But you're not going to get 10ft accuracy even averaging over 2 minutes because some of the error sources are systematic rather than random, at least over a short timescale. Professional use either uses much bulkier more expensive units often combined with differential GPS
    – Chris H
    May 14, 2017 at 7:33
  • 1
    A more practical explanation is at gpsinformation.org/dale/dgps.htm - to summarise, a dedicated unit with a beacon receiver might help you.
    – Chris H
    May 14, 2017 at 10:40
  • 1
    This might be on topic at gis.stackechange; if so you're more likely to find an expert answer there. But what sort of area are you thinking of? What's the application?
    – Chris H
    May 14, 2017 at 10:47
  • 2
    My specific application is the marking and recording of buried drainage tile locations in farm fields--I need it to be accurate enough that I can find the ditch quickly in the event that I need to make repairs or modifications. However, I am also simply curious about the accuracy of GPS in general. May 14, 2017 at 13:16
  • 4
    Bury a piece of rebar (cheap) with the objects of interest. Use GPS to get close. Use metal detector to find it.
    – Jon Custer
    May 14, 2017 at 16:42

5 Answers 5

14
  1. No. GPS is not autonomous. Errors in satellite vehicles can and have propagated to ground stations (I have one on record from a working GPS receiver placing me kilometers from my actual location that coincided with a published failed software update in the satellites). Additionally, selective availability, while not likely, can be turned on by DoD any time it likes. Because there are so many thing to go wrong, it's not possible to use GPS in systems that have a hard performance requirement.

  2. Iphone (like all cell phones) has a small GPS antenna and the GPS is akin to "would you like fries with that". i.e. GPS is an add on rather than its reason for existing, and is inferior in nearly every way to a dedicated GPS unit. .

  3. Can't say. GPS results are statistical, not hard and fast. Study standard deviations and such like to get an idea of how to specify required GPS performance. After that, I am sure money comes into it, and for the right amount, you can get what you are after, but only as long as you can specify it.

You may be interested in studying up on "Receiver autonomous integrity monitoring" (RAIM) as well.

If you can't be bothered with all the above and want a simple answer - buy a dedicated GPS with the smallest TTFF (Cold start, non assisted) you can afford. This is as good a guide to sensitivity and quality as you will get without becoming a rocket scientist.

4
  • 4
    SA is supposedly permanently disabled. But it's in the hands of the US government and I'd be surprised if they didn't have the physical capability to turn it back on. Politically it would be crazy given how much had come to rely on it in the couple of decades since the switch off. Unfortunately we can't rule out crazy.
    – Chris H
    May 14, 2017 at 10:41
  • But cold TTFF on any device will be over 12 minutes, right? Maybe hot TTSF is more relevant?
    – Roflo
    May 15, 2017 at 15:23
  • 1
    Do you have hard evidence that dedicated consumer units are more accurate than a modern smartphone? I've always assumed this would be the case, but the practical experience of many walkers I have asked about this indicates otherwise. Here's a fairly typical test which suggests that phones perform pretty well: singletracks.com/blog/gps/… May 16, 2017 at 10:52
  • 1
    Accuracy is never going to be a problem with a modern GPS in an open feild. Where it makes a difference is when the GPS receiver is in fringe reception conditions- under a dense canopy is the typical concern for an outdoor GPS. Everything else being equal, the device with the bigger antenna will perform better in fringe conditions. Think cameras - you phone camera will take pretty good photos 99% of the time. When the light gets low and action gets fast, your gonna need a bigger lens so you can capture more light and a DLSR will walk all over a phone. .
    – user5330
    Dec 8, 2017 at 0:01
6

Over a small area you may be able to do something. I don't know if you have a small area to deal with, and 10ft might be pushing it even then. Basically you fix the position of a known point by some means (it might not even have to be fixed absolutely depending on your goal). Long term GPS averaging might do that for you. Then you use this known value to calibrate an offset. You'd need a laptop connected to a dedicated GPS unit and probably custom code. If you could also access differential GPS signals you'd need a beacon receiver for that, but wouldn't need the known point or your own code.

But over a sufficiently small area compass triangulation can get surprisingly good relative accuracy; once again you need well-surveyed references (and line of sight to them).

2
  • The farm on which I plan to use this device fits entirely into a circle with a radius of 1 mile. Is that considered a small area or a large one? May 14, 2017 at 13:19
  • 2
    That's a small area in terms of both DGPS and triangulation, at least if there are things like chimneys, grain silos or water towers that can be seen from all points. There are farms in the US and in the UK that use DGPS-controlled machinery to improve efficiency by reducing overlap in working their fields. Maybe that would help somehow.
    – Chris H
    May 14, 2017 at 15:04
5

There are several ways to use GPS signals that would give you the desired accuracy. You can use either DGPS or a SBAS (WAAS in the US) for an accuracy of meters. For a centimeter level accuracy you would need an RTK o PPP device.

I doubt you can get the desired accuracy in less than two minutes since the startup of the device unless you have a data connection to obtain the GPS almanac (I guess you could use a satellite Internet service). Once the device has located the GPS satellites, it should be pretty fast.

In a farm setting a SBAS system is usually the most cost-effective option.

2

Ten feet isn't unreasonable, although it's right at the edge of day to day accuracy. I used a Garmin 60c along with aerial photographs to create an orienteering map. I had an overall accuracy of about 5 meters, coupled with a 8 meter differential between photo and GPS caused by the 40 meter difference in elevation between floodplain and highlands.

Ten feet from a cold start in under two minutes isn't reasonable.

A: Enable WAAS. Some of the error in GPS is due to ionization state of the upper atmosphere. (More ions, slower radio) This tends to be a large scale phenomena so by comparing the calculated GPS position, to a real location, you can figure out a correction. This correction is broadcast. This routinely gets errors down to under 3 meters -- 10 feet.

B: I found when using a GPS for making trail and cross country maps that it would take about half an hour to really lock in. The longer it had been since I last used the GPS the longer it took. If it had been off only a day, it took only a few minutes. I also found in woods that sometimes moving just a few feet would improve my accuracy. I suspect line of sight to the optimum satellite constellation was blocked by a tree.

C: The constellation of satellites in the sky makes a difference. Bunched satellites or too close to the horizon are worse than having a wide spread set over 30 degrees above the horizon.

D: Some of the errors vary over long periods (~1 hour) So the relative error of two fixes made close together in time can be quite small. I used this to good effect by having locations where I could get an accurate map fix (Benchmarks, road intersections, power pylons) Go to the location, take a waypoint, go to my new spot, take a waypoint. The difference was repeatable to 1-2 meters, if the time difference was under 10 minutes. (For orienteering having small local relative errors is more important than absolute accuracy. You want to know which side of the road a landmark is.)

E: Remember that the error is statistical, and has a Poisson distribution (long tail) When the GPS error is 3 meters, that means that half of the time, the difference between measure and actual is 3 meters or less. The other half of the time it can be more.

F: In mountains and urban locations, you can get reflections. These can move your fix by a couple of kilometers. Conditions for this in the mountains aren't common, but if your fix makes no sense, wait a few minutes.

G: A caution: Most of the US Topo maps and Canadian topo maps have contour information derived from stereo aerial photographs. The Canadian maps can have a horizontal absolute error in contour line placement of up to 90 meters, north/south and half that east west. This can result in the GPS telling you you are at the top of the cliff when you plot your location, while the reality places you at the bottom. Benchmark placement, however is very accurate.

H: Forget the phone. Fine for finding street routes, not much else.

See this thread on Robotics.SE: It talks about various differential GPS techniques that can get sub 1 meter accuracy.

https://robotics.stackexchange.com/questions/4312/low-cost-centimeter-accurate-satellite-positioning-gnss-gps

More info:

https://en.wikipedia.org/wiki/Global_Positioning_System

-3

This answer is based strictly upon my personal experience. I am not trying to sell you anything.

There are consumer grade GPS receivers which will get you approximately 3 meters of repeatable accuracy if they are used properly, and there is little or no overhead canopy.

I have had amazing results using a Garmin GLO which receives single frequency GPS, and Glonass data, along with WAAS corrections. ( About $110.00 ) It connected via bluetooth, and worked with multiple GIS, and mapping software applications, as well as various versions of the Windows operating system. I have also used it with several models of semi-rugged, and rugged tablets without issue. I have read that it is compatible with Android, Linux, IOS, and just about everything else out there. I only have Windows experience with it.

There were the occasional flyers that were in the five meter range, but in most cases, I was locating previously surveyed locations (Data collected with Survey Grade, Dual Frequency, RTK corrected, GPS/Glonass receivers) within a sub two meter radius.

One of the things I have found is, the units tend to work better when you are moving slightly. Just standing still, the position tends to walk around creating about a four meter circle.

The Garmin I was using had a 10 HZ update rate which created a very smooth line if I was mapping access routes, or roads.

The mapping data I collected, overlaid extremely well on previously located features.

There are other consumer grade external receivers that are probably capable of collecting data with this type of accuracy, I am just not experienced with them.

I have yet to see an embedded, discreet GPS chip within a mobile phone that approaches the results I have gotten with the Garmin.

Selective Availability has been turned off for a long time now, but it can be turned on at the whim of the Department of Defense. It would really serve no purpose to do so at this point in time, or in any foreseeable scenario with regards to national security. Everyone knows where everything is. (Sorry, bad generalization.)

With regards to locating drainage tiles, I have done a great deal of this using a tool called a Magnetomatic. It is basically a telescopic car antenna attached to a swivel that is mounted through a handle. When you cross something that is buried, the antenna will swing in the direction of the object or line that is buried. It works on plastic, poly, tile, transite, and metal lines, as well as fiber optic, and twisted pair copper lines. At least it does for me, and most of the people I work with. Some people are unable to get it to work for them. Most of them are left handed, and the others are people who are unable to wear a movement style watch. (I believe these people have reversed polarity, but I cannot prove it.) The process is similar to witching (divining) for water. You can also bend two pinflags at a 90 degree angle to make short handles for each, put one in each hand, with the long ends parallel to each other, and parallel to the ground, about waist high, keeping them a shoulder's width apart. They will cross when you encountered something buried.

If you are using the GPS to measure a distance from a known location to find other unknown locations, say starting at a pump in the corner of a field, wanting to find the grid of drain pipes every 100 feet, the Garmin will get you close, but you will still want to verify the locations somehow, rather than blindly trusting the locations provided by the Garmin.

3
  • 2
    This was a good answer until you started to advocate dowsing.
    – Mark
    May 14, 2017 at 19:29
  • It was an okay anecdotal answer, up until the dowsing nonsense. But even enecdotally it doesn't actually give any useful information to the OP. Consumer grade GPS just categorically can not give the accuracy required.
    – Rory Alsop
    May 14, 2017 at 21:53
  • I do not know a lot about GPS, but I just did some quick research and found that, contrary to what others have said, consumer grade GPS can get the accuracy required. However, to simply say "Yeah, they'll do that" is an incomplete answer; this answer glosses over some of the points that make its specific receiver capable of what is claimed in it, but it does mention a setup that is capable of "under 3 meters" according to gpsinformation.org. I'm very skeptical of the divining section, but this answer is still useful. +1
    – Loduwijk
    Sep 6, 2018 at 20:28

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.