I think what may be happening is that the GPS is averaging numerous measurements, each within a few foot radius, to get your approximate spot. Its a nifty trick to improve the instantaneous accuracy of the unit :).
When you move along a straight line, the average point occasionally jigs away from your correct spot. Should you stand still on a spot long enough, the fluctuations between readings will converge, until it tells you your exact position on earth down to the hundredth of an inch (or better)!
As you’ve surmised, drawing the imagery on the maps is a tricky business. You’ve actually gotten it slightly backwards, as its the position given in latitude and longitude that is the most accurate (down to the hundredth of a foot). Aligning the aerial photography with longitude and latitude is the real challenge! Due to distortions caused by the curvature of the earth, or the angle the photo was taken, imagery alignment can range from spot on, to off by a few feet.
Modern maps such as provided by Google Earth generally have well aligned photography, but older imagery should be taken with a grain of salt. That your GPS shows your position accurately relative to a random tree in the woods is a testament to Google’s process of orthogonal rectification!
Civilian standard household units/equipment, like your SatNav or phone, has an accuracy of about 3m.
However, GPS can be as accurate as 10-30cm (4-12inch) with special equipment; military and special permitted construction etc.
With CPGPS it’s even as close as 3cm (1.2inch)
By a lovely coincidence, I just read John Keay’s “The Great Arc,” a brief history of Lambton and Everest’s “Great Trigonometrical Survey” of India, which, among other things, established the heights of the Himalayan peaks. The profession of surveying has certainly been changed by GPS sats!
I just taught a class about surveying. On the opening day, we were talking about old methods of measurement.
The students were so confused about converting 66 links in a chain, 4 rods to a chain, 16.5 feet per rod, etc. I finally dug an old surveying chain out of a closet to show them what the heck the book was talking about!
I’ve never found any phone to map my home location adequately. This came up recently because my newspaper delivery was continually going to the house across the street.
Regardless of the phone (or mapping service) my home gets mapped across the street.
I had to send a picture of my house and threaten to cancel my subscription for the newspaper circulation department to get the delivery person to look at the addresses rather than their phone.
a chain is 66 feet long, which is 100 links. chains are made up of 100 links which are each 0.201168m long.
i’ve not heard of a chain made up of foot long lengths.
to the OP, i’ve generally found that when average to good satellite configurations are overhead, and there are few obstructions, then i expect the raw WGS84 positions reported by the smartphones i’ve used to usually be good to under 5m. i’ve not used a dedicated handheld GPS for a while but the phones seem to be just as good as what i was using some years ago.
you may experience an apparent inaccuracy if you move, and your position does not update quickly enough. you will be seeing an older position being reported for a while before you receive the updated position.
when viewing your position overlaid on a map showing aerial imagery, for example, your ‘good’ position may appear to be inaccurate due to the error in the georeferencing of the underlying imagery; your GPS may be calculating an accurate fix, but it will look out of position with respect to the (inaccurately located) features shown on the map.
Do you mean that if you type the address into a mapping app then it comes up with the wrong house, or do you mean that the “You Are Here” dot shows you as being in the house across the street?
The former is very common and had nothing to do with GPS accuracy - it’s simply down to inaccurate geocoding in the map data. Millions of addresses are inaccurately mapped - many streets seem not to have any house number locations fixed on the map so the best you can get is a vague location. Here in the UK it is very much the exception rather than the rule to get the right house coming up in Google Street View if you type in the house number.
If it’s the latter (which would be much less common) then that is far more likely to be a problem with how the map is drawn and/or how the aerial photos have been stitched together. In other words, the GPS is giving accurate co-ordinates but the street is in the wrong place on the map.
A lot of people seem to think their GPS is inaccurate if it shows their house in the wrong place when they type the address, but that has nothing to do with GPS.
I think a lot of this goes back to what your local governments have available for map makers. If I go to Google maps and type in my address, it hits it dead on with the correct street view and everything. It even shows my property lines in the map view. That is because my county has all of this information in a way that Google can access it.
Yes, I’ve noticed how much the maps vary. If you go to the “map” view (not satellite view) then some areas show the property lines and a very accurate outline of each building. Most areas of Australia seem to actually show the property number superimposed on each lot, too (although building outlines generally aren’t shown). In the UK, only certain building outlines are shown (usually large landmark and/or public buildings in cities) and street numbering seems to be totally absent.
If I type in my full address, it drops a marker four houses down and on the opposite side of the street from my house.
Australia, I believe, has a Torrens systems, which means that the deed on file in the county office is a good as law with respect to property ownership. They likely have much more complete property records than in the United States.
It also helps many landrecords date back only to the 20th century, whereas in the United States, many areas were settled in the 1600’s and have a mess of probate issues.
I used to play with GPS, doing experiments and whatnot. IIRC I found a standard deviation of repeat measurements of a location something like 40 feet using my iPhone 4S. I have several Garmin handhelds (I think the model was GPS-60) that can do “waypoint averaging”, meaning they average the position information for as long as you want to wait. I found the standard deviation of repeat measurements dropped to 1.5 inches when I let the averaging time go to two weeks. I’d use a DC power supply, and an external antenna on my roof. I also figured out that Garmin alters the last few digits, which would normally be insignificant, for some reason when moving waypoints in and out of the device in some circumstances, and suspect that they did it as a cheat to let them encode something like symbol color for waypoints as an afterthought.
The biggest difference with surveying GPS is that they keep a separate record of the pseudorange from the receiver to each satellite separately in the data record, and let you use a reference station to correct distance errors on a per-satellite basis. They also use multiple periodic things, both the radio frequency carrier wave, and the repeat time of the pseudorandom code stream.
GPS is amazing. There’s a saying that you should “keep it simple, stupid!” but GPS is a monument to just what an amazing thing can be accomplished with a system that is really, really complicated.
TLDR : GPS is nowhere near this accurate normally. The reason is that there are unknown (to your receiver) factors in the atmosphere that affect how long the signal takes to reach the receiver).
This puts a hard limit on accuracy, normally. More satellites lets you narrow the range down, and using the satellites run by other countries (the russians have a system, the EU has a few) can also help.
There is a fix for the atmospheric distortion problem : you need to have a base station in a known location that receives the distorted GPS signals, calculates the base stations position, compares it to it’s actual position, and broadcasts a correction message. This is called differential GPS.
There are varying levels of dGPS. Some of it is done on a very broad scale, with just a couple of base stations per state, and a satellite signal broadcast to a large area. This works almost anywhere, but is not as accurate. In densely populated areas, there might be more stations.
Anyways, it’s a complex subject and it depends on a lot of factors, including, crucially, which differential signals your receive has access to.
Wait, you mean I can’t do property surveys with my super accurate find me on my couch phone? Well, that is no fair. I bet the survey companies are lobbying like crazy to keep that knowledge a secret.
Of course, that was the 3G, a couple of generations back. An empirical test showed quite good accuracy compared to dedicated GPS units: Informal GPS Logger Test: iPhone 4s GPS is Shockingly Good. But then again, that’s just one guy.
Stack Overflow responses seem to indicate that the chip itself has down to 10 m accuracy, but could be worse than that in practice, depending on how well the app developer deals with the hardware capabilities. Coupled with the inertial sensor, if you provided a way to calibrate it manually to a known accurate point, you might get better than that, but I wouldn’t try to use it to steer an autonomous craft or something ambitious like that.
Yes, that’s the point I was making. But lots of people blame “GPS” for this kind of error, when it is nothing of the sort.
I have an iPhone 5 (and a 4 before that) and use ViewRanger to record my routes when I’m out hiking or cycling. When I overlay the GPS track with Google Earth, the track is almost always right on the visible path. The only time it seems to be less accurate is in dense tree cover, and when corners get “cut off” (which is more a function of how frequently the app records location points, rather than a problem with accuracy per se - it just draws a straight line between each point it records, so if I’m travelling fast and turn a corner then it can miss a few metres of my track).
