Accuracy stored with waypoints?

[+Geo-Actuary]

Hi,

 

Does anyone ahve a unit that stores the accuracy of the reading with the waypoint when it records it. I was thinking that that would be useful.

 

I have a GPSMAP76S and contacted Garmin about this. They said that the unit can't do it, but they woulod add it as a suggestion.

 

Has anyone else wondered about this?

 

Thanks.

 

Nothing Worth Knowing Reveals Its Secrets Easily - A Friend

[Kerry.]

But what really will say garmin accuracy mean to someone with another unit as any of this estimated accuracy/error stuff is simply that, estimated and relative to the method a particular manufacturer implements their own "accuracy estimation", and all manufacturers do that differently and unless all the different manufacturers are going to release "their" method then it basically means absolutely ziltch.

 

Cheers, Kerry.

 

I never get lost everybody keeps telling me where to go

[+phantom4099]

EPE is put in there so we know who the new people are.

 

Wyatt W.

 

The probability of someone watching you is directly proportional to the stupidity of your actions.

[+Searching_ut]

It's pretty rare that I look at the EPE readings my receiver gives me. I tend to make my accuracy guesses based on stability of the position/altitude fix the unit is giving me.

 

Since you're using a Garmin unit, one way you can save and utilize HDOP PDOP etc is to hook up your receiver to a PDA and use a program like VisualGPS to average your waypoints, save the NMEA data etc. You can even set up thresholds on the software so that in theory, it will only use data with a fairly high probability of accuracy in the averaging. Personally, I've found it to be fun to play with, but I don't ever tend to actually use it for anything other than playing around.

[+Geo-Actuary]

Hi Kerry,

 

Using that line of reasoning, each GPS unit made by each manufacturer would produce results, not only in the accuracy measurement, but in the basic measurement of Lat/Lon and all other readings, that would not be at all compatible with other units. But, ... we all share waypoints that we record and we all are able to find them using different GPS units.

 

If someone records a location and has an accuracy of 30' and you locate their waypoint with an accuracy of 30', you could easily be standing right on top of it or you could be 60' away from it. All I'm suggesting is that the unit's ability to estimate accuracy is reasonable, no matter which unit it is, since they all use the same satellite readings and the same level of mathematics to determine distance and timing information. See my response to 'Searching-ut' for another view of the use of the accuracy field.

 

Mike

 

quote:

---

Originally posted by Kerry:

But what really will say garmin accuracy mean to someone with another unit as any of this estimated accuracy/error stuff is simply that, estimated and relative to the method a particular manufacturer implements their own "accuracy estimation", and all manufacturers do that differently and unless all the different manufacturers are going to release "their" method then it basically means absolutely ziltch.

 

Cheers, Kerry.

 

I never get lost everybody keeps telling me where to go

---

 

Nothing Worth Knowing Reveals Its Secrets Easily - A Friend

 

[This message was edited by Geo-Actuary on May 24, 2003 at 04:56 AM.]

[+Geo-Actuary]

Hi phantom4099,

 

Thanks for the helpful reply. Sorry we all aren't up to your standards.

 

quote:

---

Originally posted by phantom4099:

EPE is put in there so we know who the new people are.

 

Wyatt W.

 

The probability of someone watching you is directly proportional to the stupidity of your actions.

---

 

Nothing Worth Knowing Reveals Its Secrets Easily - A Friend

[+Geo-Actuary]

Hi Searching_ut,

 

What you are really saying here is the following:

 

1- You rely on the accuracy of all the other mathematical calculations that are produced by the GPS and displayed on the screen,

 

2- you rely on your ability to quickly analyze and track the minute changes in the position/altitude fix, and

 

3- you then arrive at a feeling of the accuracy of your unit based on some mental model that you apply consistently while your out moving around the field.

 

How is that better than the what the unit itself can do? After all, the unit doesn't make any guesses, it records the same information that you use (changes in the fix/altitude) at a faster and more accurate rate than you, it applies the information it gathers consistently from point to point, and it is not distracted by bugs, terrain, lack of sleep, not enough coffee, etc.

 

By the way, thanks for the info on the PDA approach.

 

quote:

---

Originally posted by Searching_ut:

It's pretty rare that I look at the EPE readings my receiver gives me. I tend to make my accuracy guesses based on stability of the position/altitude fix the unit is giving me.

 

Since you're using a Garmin unit, one way you can save and utilize HDOP PDOP etc is to hook up your receiver to a PDA and use a program like VisualGPS to average your waypoints, save the NMEA data etc. You can even set up thresholds on the software so that in theory, it will only use data with a fairly high probability of accuracy in the averaging. Personally, I've found it to be fun to play with, but I don't ever tend to actually use it for anything other than playing around.

---

 

Nothing Worth Knowing Reveals Its Secrets Easily - A Friend

 

[This message was edited by Geo-Actuary on May 24, 2003 at 05:07 AM.]

[+Sissy-n-CR]

Should have bought a SporTrak Map and you wouldn't have to worry about inaccuracies.

 

Just kidding.

 

I think mostly what the guys are saying is, in a nut shell, unless all the manufacturers of GPSs use the same math to determine position and possible error then the number would be practically useless. You can guess-timate your accuracy by looking at the satellite positions screen, but that only tells you your accuracy.

 

Then you have new users or users who don't care about accuracy putting up coords and even the best system is useless. There is always the human element and it will cause errors.

 

As for the smartassed remark, I'm sure it's just fun being poked. Most likely everyone has gone through the phase of wanting more and better accuracy, feeling their unit is better than everyone else's, and/or this site would be just this much better if it added this or that feature. Soon, you start to realize a lot of what you thought was important in the beginning, isn't. You'll start worrying about other things like which PDA to use or what software is the best for your setup. You might even get hard and heavy into the forums, become an "expert," give authoritative opinions on a host of issues, and then slowly fade away. Hell, I've only been around for a few months and I've gone through it and seen it in others.

 

"More accuracy" is a typical line of thought in new cachers. Everyone has their own opinion of what can make things better. Some come up with a better way, some think of the same thing a host of others have thought up and been shown why it won't work.

 

Pretty soon you'll probably realize complete accuracy between the posted coords and your reading on top of a cache is pretty much not possible with today's consumer equipment--from my understanding, even the $50,000 high-end professional models can't give you on-the-spot readings in real-time on the field.

 

Then you have to ask yourself the question, is making your unit say zero and you bumping your toe on the cache really what you want? Most with more than a few dozen finds would say "no."

 

It's nice to wonder about different aspects of the game, but it just appears that you're in a particular phase that most of us have gone through. I've gone from posting coords with my finds if they were a couple of points off to not worrying about it unless they're 100'+ off and even then I only post an approximate distance and direction to help the next person narrow it down.

 

This is all just my opinion as I don't want to be accused of being one of the "resident blowhards" (I probably am, anyway), I just wanted to say what I thought on this topic.

 

Welcome to our obsession.

 

CR

 

Oh, one other thing, you can put more weight on what Kerry says about GPSs that most of the rest of us. There are a few in these forums who are true experts on GPSs and the technology behind it. I apologize that I can't name names, but my coffee hasn't kicked yet in this morning.

 

[+Geo-Actuary]

Hi CR,

 

I agree with much of what you say about human error and the practical use of the accuracy readings. I even think that your right about being able to stand right on top of the cache with a zero reading - that would take some of the fun out of it.

 

It's not a big deal, really, my only point was that it would be nice to see the accuracy that your unit had when it took the reading. For those who don't believe it's useful, they could turn it off. For others, it might be helpful.

 

As you know, the accuracy is affected by many things, weather, surrounding terrain, leaves on trees, and even the unit itself. I don't believe the algorithms that are used by the various manufacturers are really that different. Based on the data that their units receive, they probably use fairly standard formulas to calculate the error - especially if WAAS is active. So, if someone goes out on a clear day in the spring with no leaves overhead, and hides a cache, they may get a reading of 10 feet accuracy. Then I go out in the summer on a cloudy day looking for the cache and I get a reading of 50 ft accuracy. I know that I need to look in an area of about 60-70 feet. If I go out a day after they hide it under the same conditions and I get a 10 foot reading, then I am probably within 20 feet of the cache. No big deal, just more data.

 

Mike

 

quote:

---

I think mostly what the guys are saying is, in a nut shell, unless all the manufacturers of GPSs use the _same_ math to determine position and _possible error_ then the number would be practically useless. You can guess-timate your accuracy by looking at the satellite positions screen, but that only tells you _your_ accuracy.

---

 

Nothing Worth Knowing Reveals Its Secrets Easily - A Friend

[+Sissy-n-CR]

I hear you. You like to play with numbers. That's cool.

 

You'd probably have more success getting manufacturers adding downloadable error data than getting hiders to put useful hints on their pages!

 

"I'm a pine cone under some pine needles." Now that's a useful hint when you're standing in a pine forest! LOL

 

I've got one that some might consider useless, but if you know how I think, it'd be obvious. However, if you exhaust the clue, you'd be just beyond arm's reach of the cache and be able to see it. This one is just a little more obscure than the rest.

 

Anyway, you'll find coordinates so far off that they are outside even the worse error. I've seen several 100' plus. some caches are reported time and again to be off and the owner does nothing. There's one that skunked us one night where the coords placed us in the middle of a road. Others have reported the same thing. The owner has yet to adjust the coords.

 

That's something you'll soon learn to live with.

 

Getting people to take several readings and post their average (especially is less than ideal conditions), doing the same on each maintenence trip, accurate cache descriptions, and meaningful hints, all would go further to having a better/easier time at the endgame.

 

But that's all part of the challenge! If it was easy, we'd soon get bored and find something else.

 

CR

 

[+Cachetrotters]

Perhaps a quick primer on the metrology of the word accuracy we bandy about around here is in order.

 

Accuracy is the closeness of agreement between a measured value and the accepted or true value. In our case, the position displayed on our GPSr and the published coordinates for a benchmark, perhaps. A GPS surveyed BM even better? That difference is the measurement error.

 

Precision is the closeness of agreement between repeated measurements under similar conditions. What does your GPSr indicate when you return to the same BM time and again.

 

The interesting thing about accuracy is that it is a qualitative statement. You can be very accurate, but you cannot be accurate to 30 feet. That is an error statement. How much error you can tolerate describes the accuracy.

 

So, saying "My GPS is more accurate than yours" is perfectly fine, as long as you accept the response "No, I think your quality standards are low, therefor my GPS is more accurate than yours". No numbers are required, only twisted geo-minds engaged in verbal one-upmanship.

 

Precision on the other hand is a quantitative statement resulting from random errors. You can measure precision with numbers and evaluate it with statistical techniques. Which is how the mfg. comes up with the "Accurate to <3meters at 2drms 95%" babble. That is really an error statement (combined random and systematic error) with the required interval and level of confidence that characterizes the error and gives it meaning. We don't hear "My GPS is more precise than yours" very often around here. That would be more meaningful, but certainly less exciting.

 

Interestingly, most all consumer-grade GPSr's have the same error statements. Hhmmm.... Should I suspect most of the error in positioning originates from everywhere except the receiver? Looks like any error introduced by the receiver is down in the noise (aside from gross errors).

 

Hey, Sissy-n-CR, I think this fills the blowhard requirement. Where do I apply to be a resident?

 

don--

up-and-coming blowhard....

[+Sissy-n-CR]

quote:

---

Originally posted by Cachetrotters:

Hey, _Sissy-n-CR_, I think this fills the blowhard requirement. Where do I apply to be a resident?

---

 

Owich! You made my brain hurt. ~rubs temples~

 

I think that makes you a member automatically!

 

CR

 

[+Cachetrotters]

No membership fee. Cool.

[+Searching_ut]

quote:

---

1- You rely on the accuracy of all the other mathematical calculations that are produced by the GPS and displayed on the screen

---

 

I don’t rely solely on the GPS for navigational data period. It’s a tool I use, but know and understand is prone to error as I’ve mentioned on my webpage. I still have more faith in my map, compass, and basic navigations skills.

 

quote:

---

How is that better than the what the unit itself can do? After all, the unit doesn't make any guesses, it records the same information that you use (changes in the fix/altitude) at a faster and more accurate rate than you, it applies the information it gathers consistently from point to point, and it is not distracted by bugs, terrain, lack of sleep, not enough coffee, etc.

---

 

I think the theory of the computer being able to make better computations is a good one, the problem is it doesn’t take much of the available data into the equation. If you e-mail Garmin or Magellan asking how the compute EPE, or what percentage of the time the formulas should have you inside that value, they won’t tell you. “Trade secrets”. Play with the units in the field a little bit though, and you’ll quickly learn it’s most likely a simple satellite geometry equation. The EPE can be a fairly good guess on flat level ground with little to no tree cover. Head to a tree filled canyon where multipath comes into play, and the EPE numbers become somewhat useless. You’ll often get EPE’s of 30 feet or less while the actual error is hundreds of feet. In doing side by side receiver tests where I was keeping a close eye on things like EPE when trying to return to saved waypoints etc. My own conclusions were that EPE numbers aren’t a particularly useful tool in the reception problem areas, which is where you actually want and need some input.

 

The end result is that my now about 8 years experience with handheld receivers, covering thousands of trail miles I really do think I can make better guesses at what sort of accuracy the receiver will give me that it does. The limitations of the EPE numbers become especially apparent when you use and compare receivers side by side. How can two receivers who both claim EPE inside 20 feet be reporting positions 80 feet from one another?

 

As for actually quantifiably measuring the accuracy of the EPE numbers, that isn't something I've seen anyone do anywhere. If you have a lot of benchmarks around, in a combination of good and bad reception areas, doing actual measurments would be something many would like to see the data on. Based on my own measurments of repeatability in reception problem areas, I'd guess the EPE numbers to be accurate less than 50 percent of the time in tough conditions. It would be interesting to see it actually quantified via true accuracy measurments however.

[Kerry.]

quote:

---

Originally posted by Geo-Actuary:

Hi Kerry,

 

Using that line of reasoning, each GPS unit made by each manufacturer would produce results, not only in the accuracy measurement, but in the basic measurement of Lat/Lon and all other readings, that would not be at all compatible with other units. But, ... we all share waypoints that we record and we all are able to find them using different GPS units.

 

If someone records a location and has an accuracy of 30' and you locate their waypoint with an accuracy of 30', you could easily be standing right on top of it or you could be 60' away from it. All I'm suggesting is that the unit's ability to estimate accuracy is reasonable, no matter which unit it is, since they all use the same satellite readings and the same level of mathematics to determine distance and timing information. See my response to 'Searching-ut' for another view of the use of the accuracy field.

 

Mike

---

 

Mike, as many of the others have mentioned EPE is one of those things that one really shouldn't put all that much reliability in. Now if the manufacturers really wanted it to mean something then they'd either release the actual method or all adopt a "standard" (how about DOP's )

 

Even the manufacturers have/had issues with (system) accuracy statements, statistics (important word that, statistics) and to make it all sound better they invented the CEP spec. CEP, RMS 95% etc, why not simply accept the same accuracy spec from the same satellites that use the same principles.

 

When one compares the old accuracy spec to the new, one of the main variables that is missing is the "user factor". Even if a unit was reasonbable (and they basically are BUT within reason) the user is a very uncontrollable factor .

 

Many of the issues with estimated accuracy were highlighted when SA was set to zero as it became very evident very quickly that the manufacturers were "cooking" the numbers. Frankly they simply didn't reflect the real world and many were caught out. These days there's really not a lot of room to move and if one analizes EPE's against "actuals" one will generally find that the lows and the highs basically never show on the display. Again one could think the manufacturers are "fiddling" the numbers.

 

Cheers, Kerry.

 

I never get lost everybody keeps telling me where to go

[+apersson850]

I think it's rather obvious that they are "fiddling". After one update of firmware for my Vista (don't remember the release number now), my Vista suddenly started to show consistently smaller EPE values. I didn't have any GPS when SA was on, so it has nothing to do with that.

 

So, how come? Well, I don't know, but I can guess, with some accuracy, or could it be precision? How about this scenario:

"Market people says that our customers are complaining. The accuracy of our units is inferior to units from the competitor X. How do they know? Do they not reach the same position in the real world, when they are looking for the same waypoint? Oh yes they do, but the EPE value on our receiver says 13 meters, while X's units claim 7 meters. How can we make our units more accurate? Well, they are as accurate, but we just aren't that optimistic in our EPE calculation. OK, then it's easy. X's receivers report a EPE that's half of what we have. Let's take our previous value and divide it by 2.5. Then well come out on top!"

 

I can add that this is completely fictious, but I wouldn't be surprised if a similar discussion actually has occured.

 

Now, it could of course be beneficiary if the caches we are looking for had the EPE published with the coordinates. But I agree with you, Kerry, that it isn't really meaningful when the EPE is calculated diffently by different units. If you go there with an X receiver, and get an EPE of 11 meters, it may be that an Y receiver would have displayed that as 7 meters. That's an area 2.5 times larger to search. Publishing something that isn't up to market departements desire for the units to "look" good, like the PDOP, would be better, but units like my Vista doesn't tell you any DOP value, unless you connect something to it and listen to the NMEA output.

 

Anders

[Kerry.]

THe following are actual EPE numbers verses actual accuracy. 3 different models, all the same brand but range over the years since 1994 and have software versions that reflect their era. The data is based over 23 hours and the oldest one (pre SA software version) was done twice for confirmation.

 

All are straight stock standard GPS positions (no augmentation)

 

Day 1, Model X (as current as one could get)

EPE average 4.16m, range 3.3 to 6.5m

Actual accuracy @50% 1.7m (max 5.4m)

 

Day 2, Model Y (superceded model but post SA software)

EPE average 3.56m, range 2.8m to 6.4m

Actual accuracy @50% 1.7m (max 7.6m)

 

Day 3, Model Z (superceded model pre SA software)

EPE average 11.4m, range 8.9m to 101.3m

Actual accuracy @50% 3.1m (max 19.0m)

 

Day 4, Model Z

EPE average 11.4m, range 8.7m to 95.5m

Actual accuracy @50% 3.1m (max 18.1m)

 

One can ponder what those numbers actually mean but for what it's worth, my thoughts:

 

There really is bugga all difference in the actual accuracy (in practical terms) and certainly well below the system spec. Model Z only outputs to 3 significant decimals (the 2 other units capable of 4) so around 3 metres is probably the best it's going to do anyway.

 

The EPE for Model Z (pre SA software version) doesn't even come close to the current real world accuracy and follows the thinking that EPE is rather a fudge figure and doesn't follow any set rules.

 

It's also interesting that the maximum EPE for Model A (Day1) was greater then the maximum actual accuracy. Not by a great lot but all the same why should the EPE exceed the "actual" error?

 

Based on Model X & Y the EPE basically sticks to the (average) middle ground generally ignoring the low & high accuracy numbers.

 

Cheers, Kerry.

 

I never get lost everybody keeps telling me where to go

[+apersson850]

To me, it seems that the EPE indicator isn't useless, but it could just as well have been a graphic bar, without any scale marks. Short bar = better, long bar = worse.

As it's now, I think the indication will fool a lot of users into beleiving that this is fact, not fiction (OK, let's call it clever guesswork, then). In your mind, Estimated Position Error becomes just Position Error.

 

Anders