I don't want to theorize to deeply on this, but I doubt that iOS is using any single calculation standard to return a value. It's more likely a custom hybrid of a number of calculations, based on the conditions at the time of the measurement (IE WiFi/Cell/GPS availability, Timeout limit of request, etc..).
For instance, according to this: http://support.apple.com/kb/HT5467 iOS is using it's access to WiFi, GPS and Cell tower clusters to return position. Each of these networks may return different levels of accuracy on their own. A calculation based on a wifi cluster might use R95, whereas the GPS measurement might use CEP.
This article: http://www.macworld.com/article/1159528/how_iphone_location_works.html also explains that the iPhone uses A-GPS. Diving more onto wikipedia you can see where this could go into the need for custom measurements.
If I had to make a call, I would say that R95 is the general standard being used. R95 is explained here: http://www.igage.com/mp/GPSAccuracy.htm as "The radius of a circle centered at the true position, containing the position estimate with probability of 95%.", kind of like how location services are visually represented in things like "current location" on maps in iOS.
This is really interesting "rocket science" stuff though. If you DO find an answer (would be interested to see what filemaker and apple engineers had to say), please post it for us!
I have a hard time believing that calling the same function might return CEP one time and R95 the next. That would be profoundly bad library design, and I like to think Apple and FileMaker are better than that. It wouldn't be much better if the function returned an extra value indicating which error radius method it was using each time; converting between the methods is too obvious an improvement to the function.
What makes you say R95 is the favorite by convention? Your last source gives CEP that honor.
I'm sure Filemaker is just using what Apple gives, so I'm not confident that anyone at Filemaker is going to have the answers to your question.
I didn't mean to say that they are ACTUALLY using two different methods, but my imagination is that apple is using a proprietary, single base function, that is a hybrid of multiple methods. This would be needed to handle the multiple methods of which the iPhone uses A-GPS.
For example, a GPS device ONLY uses GPS satellites for triangulation, so they can rely on a single method of determining accuracy. With A-GPS devices though (most cell phones), you can also use cell towers (broken up into subsets, like CDMA, EDGE, etc..) and WiFi network broadcasts to assist as well, thus a single standard may not be ideal.
As for my R95 guess, it's just a guess based on the visual representation. On a map, you're location is notated with a blue dot in your general "true position", then there is a circle that expands and contracts, larger radius equates to less accuracy. You will see the circle constantly resize as the reading becomes more/less accurate. When you think about that visually, then the definition of R95 I quoted above just looked good to me:
"The radius of a circle centered at the true position, containing the position estimate with probability of 95%."
As noted, it's rocket-science stuff to me. I don't know enough and didn't really dive into the math enough to make a more educated guess. I'm purely speculating based on skimming of the quoted articles and my observations of personal use of location services.
I put together a file to help us try to figure this out empirically. I need everybody's help, though. Here's how it works:
- Download the file to whatever iOS devices you have with FileMaker Go (12) on them.
- Find a spot, and get comfortable. It will be helpful if you temporarily turn auto-lock off.
- Start the file recording location measurements. It will keep repeating until you tell it to stop or the battery dies.
- Leave your device alone for a long time. The longer the better. A few hours would be great, if you can manage. 30 minutes probably isn't enough.
- Repeat 2-4 in as many different spots as you can.
- Share your sample data with me. The file can export a CSV and email it to me.
Many folks are concerned about their privacy and reluctant to share data on their whereabouts, but there's no need to worry. The file will randomly offset all your data so that I have no practical clue where you've been — even if you send me the FileMaker file instead of the exported CSV. (But it does this in a way that won't upset any of the analysis it's doing to answer the question at hand.)
AccuracyTesting.fmp12.zip 154.0 K
I'm doing a blog post series on using the Location functions. Today's post is about what I've figured out so far from the location sample data I've received: