Vista WAAS Test w/USGS Benchmark (long)

[+BletchleyPark]

Yesterday, I performed an admittedly unscientific test of the WAAS feature of my Vista. The results were interesting and may indicate that the traditional "Accuracy" (in feet) figure is not necessarily the only measure of the WAAS advantage.

 

First, a little about the test. I conducted it on July 26th at 11:30 am in the clear skies on the summit of Mt. Holyoke in western Massachusetts. My GPSr is a Garmin Vista with software version 2.28; I was using freshly-charged NiMH batteries.

 

For the location of the test I chose a USGS Benchmark (MZ1645) located at the Mt. Holyoke summit. This Benchmark is located at N42°18.028'/W072°35.278', 938 ft in elevation. From this location, there is approxmately a 260° view of the horizon. The Vista was placed on the Benchmark with its patch antenna centerd on the mark and the unit supported in a level position.

 

Four tests were run, alternating between WAAS disabled/enabled. Care was taken to keep the unit undisturbed during the reconfiguration between each test. The Vista was allowed to "stabilize" for 10 minutes during each test; after which the "Accuracy" in feet and Lat/Lon position were recorded.

 

It took about an hour to perform the tests during which the satellite geometry did not "significantly" change, though there was some movement. During the tests, 8 satellites (5 above 45° elevation and 3 below) were constantly shown with signal strengths of 50% or higher. A ninth satellite located below 45° faded in and out during each of the 4 tests (being partially blocked by the Summit House structure). I was careful to stand away from the Benchmark in the "shadow" of the structure.

 

The results of the tests are as follows:

Test 1 - No WAAS

Accuracy: 20ft

Lat/Lon: 42°18.029'/072°35.274'

Test 2 - WAAS

Accuracy: 21ft

Lat/Lon: 42°18.028'/072°35.278'

Test 3 - No WAAS

Accuracy: 21ft

Lat/Lon: 42°18.031'/072°35.275'

Test 4 - WAAS

Accuracy: 18ft

Lat/Lon: 42°18.029'/072°35.278'

 

Granted that these tests are not scientific and there are only four samples, it none the less raises the following interests in my thinking:

1. Tests 2 and 4 (WAAS) were *very* close the proper Lat/Lon with Test 2 being exactly correct.

2. Both WAAS tests reported position more accurately that the No WAAS tests.

3. "Accuracy" (in feet) figures did not quite correlate with the relative positional (Lat/Lon) accuracy. In fact, for the two WAAS tests, the most accurate (Test 2 exact) Lan/Lon position reported an "Accuracy" figure worse that the least accurate (Test 4).

 

Could we (perhaps at the limit of acceptance) draw a hypothesis that WAAS improvements are not necessarily seen as a positive change in the "Accuracy" figure but, in fact, are seen in more accurate Lat/Lon position data?

 

Have any other folks done similar Benchmark WAAS testing in a more scientific environment or with a greater number of samples? It would be interesting to know of the results.

 

BP

[+BletchleyPark]

The PID of the Benchmark should be MZ1654.

 

Sorry,

 

BP

[+EraSeek]

Yes, look down the list of topics in this area and you'll see my topic on WAAS myths where we have dicussed some of this. Here is my test:

 

"A Test

 

I just got back from doing a test on top of a 'First Order Horizontal Control' NSG marker. I took a while to get a stable lock on the WAAS satellites because of the area, but when I did this is what I got:

 

With WAAS: accuracy reading of 7.7' and actual distance to mark 3.35'

 

I then switched WAAS off and got: Accuracy 18' and actual 7.3'

 

In each case about 40 pecent difference between the Accuracy reading and the actual.

 

And about a 54% reduction in actual error with WAAS."

 

We also dicussed how the ACCURACY READING or EPE is a big myth as well, as it only gives you some idea how your GPSR is doing, and that is about all.

I believe your accuracy reading is going to be dealt with differently in different GPS's and their software. Some may work better than others, but none are perfect nor should be seen so.

 

[+Alan2]

GPS Accuracy in a real geocache test

 

Here’s the result in real geocaching GPS accuracy. The set-up was that each cache finder to a NGS benchmark, cache http://www.geocaching.com/seek/cache_details.asp?ID=10814 would report their GPS coordinates at the benchmark to the cache web page. The benchmark center was on a 7 foot wide, 2 ½ foot high granite slab. You would have to reach to set the GPS in the center and to read it. Benchmark is in an open area although there are a couple of small trees nearby. The actual NGS coordinates are 40 44' 34.46640"(N) 073 50' 42.31409"(W) which places it in Flushing Meadow Park, Queens, NYC which is in Waas range.

 

The following GPS’s were reported used including my own Vista when I set up the cache. Readings were taken with 13 different GPSr’s on 12 different days between 11/25/01 and 3/24/02 at different times of the day. Weather conditions were not reported.

 

Garmin Vista (4)

Garmin etrex (2)

Garmin Summit

Eagle Expedition II

Magellan Map 330

Magellan 300

Unknown

“Cheap” Garmin

Garmin Street Pilot

 

Summary of results. Difference calculated from NGS coordinates. WGS84 coordinates in DD MM.MMM were converted to UTM using Mapsource program. Distances from benchmark calculated taking the square root of the sum of the squares of the difference between the UTM latitudes and UTM longitudes in meters and converting to feet.

 

12 readings under 12 feet

5 readings between 13-25 feet

1 reading at 40 feet (note that this was the Magellan 300 which is an older unit and reads to 2 decimal places only - about 50 feet accurate I’ve been told).

 

All were single readings except for Eagle (13 readings = 3.3 feet avg.) and one of the etrexs (7 readings = 25 feet avg.). There was one WAAS reading reported (my Vista at 3.3 feet).

 

I think the results speak mountains as to the accuracy of the equipment we are using.

 

Since there was only one Waas measurement at one meter from the benchmark, you cannot state a general determination.

 

Alan

[Kerry.]

As with any augmentation system (be it WAAS, DGPS etc) the intregrity is generally more important than the inprovement in accuracy.

 

When Selective Availability was terminated it was fairly evident at that time that EPE (and similar) in many of the recreational recievers was a massaged figure. When actual accuracy improved by a factor od 7 and EPE only by a factor of 2 then somebody was kidding someone and probably still do. With no prior announcement manufacturers sort of got caught out on that one but prior to that they had some scope (< 100m @ 95%) to hide things but not anymore [< 13m @ 95% and even less than that in the real world).

 

Also has me buggered how one can quote distances to 0.01 feet or even 0.1 feet from these types of receivers. Whose kidding who

 

Same for relationships based on lats/longs to 3 decimal minutes as there can be around 8 feet floating around in them simply based on numbers without the affects of the real world.

 

Cheers, Kerry.

 

I never get lost everybody keeps telling me where to go

[Kerry.]

As with any augmentation system (be it WAAS, DGPS etc) the intregrity is generally more important than the inprovement in accuracy.

 

When Selective Availability was terminated it was fairly evident at that time that EPE (and similar) in many of the recreational recievers was a massaged figure. When actual accuracy improved by a factor od 7 and EPE only by a factor of 2 then somebody was kidding someone and probably still do. With no prior announcement manufacturers sort of got caught out on that one but prior to that they had some scope (< 100m @ 95%) to hide things but not anymore [< 13m @ 95% and even less than that in the real world).

 

Also has me buggered how one can quote distances to 0.01 feet or even 0.1 feet from these types of receivers. Whose kidding who

 

Same for relationships based on lats/longs to 3 decimal minutes as there can be around 8 feet floating around in them simply based on numbers without the affects of the real world.

 

Cheers, Kerry.

 

I never get lost everybody keeps telling me where to go

[+EraSeek]

quote:

---

Originally posted by Kerry:

Also has me buggered how one can quote distances to 0.01 feet or even 0.1 feet from these types of receivers. Whose kidding who

 

 

Cheers, Kerry.

 

I never get lost everybody keeps telling me where to go

---

 

Just relating what my GPS's reading were my friend. I was not validating them.

 

[+EraSeek]

quote:

---

Originally posted by Kerry:

Also has me buggered how one can quote distances to 0.01 feet or even 0.1 feet from these types of receivers. Whose kidding who

 

 

Cheers, Kerry.

 

I never get lost everybody keeps telling me where to go

---

 

Just relating what my GPS's reading were my friend. I was not validating them.

 

[Kerry.]

EraSeek, And really the manufacturers should know better instead of providing this type of mythical information, which is totally not achievable with this type of equipment.

 

Like the calclator syndrome all this type of stuff needs to be kept in perpective other wise new myths are born.

 

Similar with WAAS it all needs to be put in perpective.

 

Cheers, Kerry.

 

I never get lost everybody keeps telling me where to go

[Kerry.]

EraSeek, And really the manufacturers should know better instead of providing this type of mythical information, which is totally not achievable with this type of equipment.

 

Like the calclator syndrome all this type of stuff needs to be kept in perpective other wise new myths are born.

 

Similar with WAAS it all needs to be put in perpective.

 

Cheers, Kerry.

 

I never get lost everybody keeps telling me where to go

[+EraSeek]

If I stand on a known mark, and it tells me I am 3.whatever feet away, I don't know, seems pretty darn good to me. Maybe you are not giving the system enough credit (even if Garmin may give it too much).

 

I have no idea how their software works. Educate us Kerry. Without WAAS I was about 2 meters away, with it 1 meter away. Did it just make a good guess?

 

Kerry, forgive me for asking, but you seem to have a lot of knowledge in this area, but last time I looked no finds. I take it you must work in the industry, is that correct?

 

[+EraSeek]

Something has occured to me in regard to this topic. Although your unit is limited by its position format (lat and long, UTM in meters) there is really no reason for your position not to be better defined. Your unit is not a tape measure, it is really more of a timing device. It cross references the ticks of atomic clocks. The system is limited by signal reception, software quality, and hardware. The other thing that limits it is scale. The unit will never have a problem defining your position in terms lat and longitudinal degrees, but as you lessen the scale the less confidence there will be. This is always going to be true. But there is no reason you cannot define this closer and closer. My point is, if you are defining clock ticks, and not position formats, why can't you break it down into feet, or even tenths of feet. Yes, due to signal drift, confidence is lessened, but is it not still valid. If certainty is what you require then perhaps you should up the scale to degrees only. N47d W122d . and forget any further defining. It seems to me if you can define it closer, you should. The ticks of atomic clocks are very small things.

 

[+EraSeek]

Forgive me. My thoughts tend to run in strong streams.

Perhaps we are all missing the point here. Allow me to wax metaphorical. If God exists (I think he does) then he is an absolute. A preacher might tell me what God is, and that is one thing. I may disagree and think God something else a bit different. There you have another thing. We may both be right to some degree, but we are also wrong to some degree for certain, because only God himself, the absolute, know absolutely what he is.

 

In the world of the GPS, that pinpoint on the Earth is the paragon, the perfect substance. My GPS with all it's abilities does it's best to describe it to me. It's format tells me where it thinks I am, it's EPE tells me somewhat how certain it thinks it is. But it is not the point, it is not absolutly correct. It does it's best. The point is it tells me where it thinks I am, not where I am. It is an advisor

 

In the world of religion, I will trust my judgement over the preacher. In the world of the GPS I will trust Garmin before I trust myself. It is the best tool I have at hand. (hey, but always carry that compass just in case)

 

[+apersson850]

We must remember, that there is a big difference between resolution and precision.

 

Saying that two points are 112.34634276554 meters away from each other indicates a tremendous resolution (better than one Angstrom), but if it's +/- 7 meters, all these decimals aren't much worth.

 

Connect your GPS to a computer, configure the GPS to output NMEA data strings, and run VisualGPS on the computer. Let it stay there for some hours, without moving the GPS, and see how exactly the same position is reported from the GPS. Sometimes, it seems to make rather long virtual walks, in spite of really not moving at all.

 

Anders

[Kerry.]

EraSeek, in no specific order but as I'm reading things.

 

"Give the system credit" I give the system all the credit in the world but I also realize that one can only generally get a result as good as the information that goes into that result. All recreational output is based on predicted data from single frequency receivers and modeled atmospherics (plus a few other things). Basically knowing what the signal-in-space (SIS) errors, then the affects of atmospherics and the user influences I really don't get all that impressed about accuracy claims that can't be quantified relative to the system specifications.

 

"2 metres, 1 metres away" well some of the time it could be (it is in fact) but would one risk it being that all the time, no way.

 

"Work in the industry" no I don't "work" in the industry as such but I do make use of the system basically like many others but just at varying levels of accuracy (depending). Been finding things for years actually but no not in terms of what you refer but then I really didn't think "no finds" had all that much to do with GPS discussions.

 

Precision is a funny things and really is not related to accuracy as both can be/and can not be related. Just because the position displays umpteenth number of decimals doesn't mean it's precise (or for some accurate but I don't want to get into that). At the end of the day the system controls the accuracy (generally) apart from things (methods) that can circumvent and cancel out the problematic and systematic errors.

 

GPS is certainly an aid and the depth of that aid differs depending on the approach taken in utilizing this aid (GPS). To some 10 metres is enough and for that is costs X amount, to others sub millimetre is required and that costs $$$$'s and specific methods to achieve that result but at the end of the day there's no such thing as a zero error measurement system (not to this point in time anyway).

 

Really the system has very well defined specifications but for obvious reasons manufacturers (especially in pre SA days) defined accuracy that might have sounded a lot better but in fact wasn't. Probably still applies today? Which would you buy? something with an accuracy of "less than 13m @ 95% SIS"? or "less than 3m CEP"? is there a difference?

 

For far too long there's been too many myths perpetrated and especially about things like accuracy, # of channels, averaging etc etc.

 

Cheers, Kerry.

 

I never get lost everybody keeps telling me where to go

[Kerry.]

EraSeek, in no specific order but as I'm reading things.

 

"Give the system credit" I give the system all the credit in the world but I also realize that one can only generally get a result as good as the information that goes into that result. All recreational output is based on predicted data from single frequency receivers and modeled atmospherics (plus a few other things). Basically knowing what the signal-in-space (SIS) errors, then the affects of atmospherics and the user influences I really don't get all that impressed about accuracy claims that can't be quantified relative to the system specifications.

 

"2 metres, 1 metres away" well some of the time it could be (it is in fact) but would one risk it being that all the time, no way.

 

"Work in the industry" no I don't "work" in the industry as such but I do make use of the system basically like many others but just at varying levels of accuracy (depending). Been finding things for years actually but no not in terms of what you refer but then I really didn't think "no finds" had all that much to do with GPS discussions.

 

Precision is a funny things and really is not related to accuracy as both can be/and can not be related. Just because the position displays umpteenth number of decimals doesn't mean it's precise (or for some accurate but I don't want to get into that). At the end of the day the system controls the accuracy (generally) apart from things (methods) that can circumvent and cancel out the problematic and systematic errors.

 

GPS is certainly an aid and the depth of that aid differs depending on the approach taken in utilizing this aid (GPS). To some 10 metres is enough and for that is costs X amount, to others sub millimetre is required and that costs $$$$'s and specific methods to achieve that result but at the end of the day there's no such thing as a zero error measurement system (not to this point in time anyway).

 

Really the system has very well defined specifications but for obvious reasons manufacturers (especially in pre SA days) defined accuracy that might have sounded a lot better but in fact wasn't. Probably still applies today? Which would you buy? something with an accuracy of "less than 13m @ 95% SIS"? or "less than 3m CEP"? is there a difference?

 

For far too long there's been too many myths perpetrated and especially about things like accuracy, # of channels, averaging etc etc.

 

Cheers, Kerry.

 

I never get lost everybody keeps telling me where to go

[+apersson850]

I agree with you, Kerry, on virtually every point. Still EraSeek has a point, when he says that you can count clock ticks down to very short distances. Sounds crazy, but here time isn't money, but distance. Still, remember that 1 ns is 0.3 meters, or one foot. So, it takes a very small timing error be that precise (accurate? ).

 

A very small signal deviation in space, will of course make such resolution meaningless. For all consumer units, the errors add up to more than what one nanosecond is worth. So, it's like in my example. A lot of digits, but they get lost in the noise.

 

To be really sure, you have to do something like what one US officer said on a press conference during the campaign in Afghanistan, when questioned about the whereabouts of Mr Bin Ladin:

 

"We are 100% sure, that either Mr Bin Ladin is alive, or he is dead, and that he either is in Afghanistan, or somewhere else."

 

Likewise, one can always be 100% sure, that you are where the GPS tells you, or somewhere else. It's opinion about whether you are alive or dead, well, let's just leave that...

 

Anders

[+apersson850]

I agree with you, Kerry, on virtually every point. Still EraSeek has a point, when he says that you can count clock ticks down to very short distances. Sounds crazy, but here time isn't money, but distance. Still, remember that 1 ns is 0.3 meters, or one foot. So, it takes a very small timing error be that precise (accurate? ).

 

A very small signal deviation in space, will of course make such resolution meaningless. For all consumer units, the errors add up to more than what one nanosecond is worth. So, it's like in my example. A lot of digits, but they get lost in the noise.

 

To be really sure, you have to do something like what one US officer said on a press conference during the campaign in Afghanistan, when questioned about the whereabouts of Mr Bin Ladin:

 

"We are 100% sure, that either Mr Bin Ladin is alive, or he is dead, and that he either is in Afghanistan, or somewhere else."

 

Likewise, one can always be 100% sure, that you are where the GPS tells you, or somewhere else. It's opinion about whether you are alive or dead, well, let's just leave that...

 

Anders

[+BletchleyPark]

Wow...I did not realize that my topic would produce such an esoteric and educational thread. I am fascinated by the complexity of the underlying "structure" and theory of GPSr technology. It is a joy to have the information and discussion which you have presented be available in public forum.

 

However, within the context of geocaching, I feel like we may all agree that benefit of enabling WAAS (providing signal-strength and geometry are acceptable) will most likely been seen in the Lat/Lon position being improved and not necessarily the number of feet/meters of "Accuracy" being improved.

 

Is my "reductionist" conclusion on target?

 

BP

[+apersson850]

Besides that, one must keep in mind that a geocache is not at a "known" location, but the position given is measured with yet another consumer rate GPSr (most probably).

 

Even if you have a millimeter accurate survey unit, it will not help you, if the original coordinates are 10 meters off.

 

But this will of course improve over time, provided that the augmentation systems are getting better and better, and that more and more people get receivers that can benefit from this augmentation data.

 

Anders

[Kerry.]

now are we 100% sure we are 100% sure

 

But for sure the timing is one of the critical things but then a few other "things" can affect that as well and that's what we are never quite sure about. Actually a month or so ago there were some "anomolies" that were detected in the system, which have apparently been there since the early 80's but due to operational circumstances had never been detected before.

 

Now been fixed but for a few days the ranging errors were in the order of +/- 16 metres but the user performance was still within specifications. That unknown 5% (95% confidence) obviously comes in handy at times for these man made/run/operated systems. That I'm 100% sure of

 

BP, that "reductionist" conclusion could be a possibility but being a software function the displayed estimated accuracy could be tuned in different ways. At the other end of the scale outside the WAAS ground network a WAAS receiver does know it has got a problem but again it doesn't appear to show ALL the error that actually exists.

 

Really estimates of the position error appear to fluctuate around the middle values and not the upper and lower limits (maybe they're not 100% sure ).

 

Cheers, Kerry.

 

I never get lost everybody keeps telling me where to go

[Kerry.]

now are we 100% sure we are 100% sure

 

But for sure the timing is one of the critical things but then a few other "things" can affect that as well and that's what we are never quite sure about. Actually a month or so ago there were some "anomolies" that were detected in the system, which have apparently been there since the early 80's but due to operational circumstances had never been detected before.

 

Now been fixed but for a few days the ranging errors were in the order of +/- 16 metres but the user performance was still within specifications. That unknown 5% (95% confidence) obviously comes in handy at times for these man made/run/operated systems. That I'm 100% sure of

 

BP, that "reductionist" conclusion could be a possibility but being a software function the displayed estimated accuracy could be tuned in different ways. At the other end of the scale outside the WAAS ground network a WAAS receiver does know it has got a problem but again it doesn't appear to show ALL the error that actually exists.

 

Really estimates of the position error appear to fluctuate around the middle values and not the upper and lower limits (maybe they're not 100% sure ).

 

Cheers, Kerry.

 

I never get lost everybody keeps telling me where to go

[+EraSeek]

Good discussion guys. Fascinating stuff to learn about. Thanks.

 

Improvements I know about: They are working on a third WAAS satellite that will be at a higher angle so everyone will be able to recieve the signal. In atomic clocks, they are working on the Mercury Ion clock with an unbelievable small error (not that it will affect us that much).

 

[+EraSeek]

Just to answer the original question, if my thick skull got it right, about the Accuracy reading (EPE) in regards to using WAAS. From what I've seen and understand, units vary quite a bit on how the EPE reading is arrived at. My unit, the map76, does show a relationship to the improved WAAS position in it's 'Accuracy reading', but only when I have a very good lock on WAAS. What usually happens is that when I first turn WAAS on my 'Accuracy Reading' number goes WAY up (meaning a lot worse) to an error of like 65' where it had been 18'. After a few minutes of WAAS lock it drops back down to around 18', no improvement, and often will stay there. But, if I have good exposure, it will drop down to figures I will not see without WAAS. Very low. I saw it drop the other day to 6.8'. I have not seen this without WAAS. I am sure different units will respond differently. This is just what I have seen in my unit.

[+apersson850]

That's nearly 100% sure, that different units calculate the EPE differently. Especially when it comes to different brands.

 

I've also heard about units, that can be updated (as many of them can), that has been changed when SA was turned off, since some marketing guy probably thought that they wanted a smaller error figure to brag about.

 

On the other hand, that could be an untrue story. I'm not sure, neither 100% nor X%.

 

Anders

[+apersson850]

That's nearly 100% sure, that different units calculate the EPE differently. Especially when it comes to different brands.

 

I've also heard about units, that can be updated (as many of them can), that has been changed when SA was turned off, since some marketing guy probably thought that they wanted a smaller error figure to brag about.

 

On the other hand, that could be an untrue story. I'm not sure, neither 100% nor X%.

 

Anders

[Kerry.]

Anders, no that 100% true (that manufacturers "invented" the CEP accuracy and it certainly does "sound" a lot better).

 

Really some units do provide 1m accuracy about 3% of the time but of course there's the other 97% that really counts.

 

But really once manufacturers start stating accuracy figures around 3m and below then unless they also provide precision in the output to match (like at least 4 decimal minutes) there just blowing hot air.

 

Cheers, Kerry.

 

I never get lost everybody keeps telling me where to go

[Kerry.]

Anders, no that 100% true (that manufacturers "invented" the CEP accuracy and it certainly does "sound" a lot better).

 

Really some units do provide 1m accuracy about 3% of the time but of course there's the other 97% that really counts.

 

But really once manufacturers start stating accuracy figures around 3m and below then unless they also provide precision in the output to match (like at least 4 decimal minutes) there just blowing hot air.

 

Cheers, Kerry.

 

I never get lost everybody keeps telling me where to go

[+apersson850]

Let's see, 0.001 minute must be 1.852 meters, along a longitude, and almost exactly one meter along the latitude, where I am (N 57º), right?

 

Which means that one digit only jumps halfway out of a 3 m radius circle.

 

You're right, Kerry, the air is hot

Today, it really is here, by the way. Around 28ºC already, which is a lot for this part of the world.

 

Anders

[+apersson850]

Let's see, 0.001 minute must be 1.852 meters, along a longitude, and almost exactly one meter along the latitude, where I am (N 57º), right?

 

Which means that one digit only jumps halfway out of a 3 m radius circle.

 

You're right, Kerry, the air is hot

Today, it really is here, by the way. Around 28ºC already, which is a lot for this part of the world.

 

Anders

[JDMC]

And why hasn't anyone mentioned altitude? It is a 3 dimentional problem the GPSr has to solve, isn't it?

[+apersson850]

It could be because the main interest among people looking at this site is geocaching (wild guess here). Geocaching is, by nature, usually done on the ground. Determining the altitude isn't that important, then. It so to speak solves itself, when you walk on terra firma.

 

But for the GPS itself it's important, of course, since no position is complete without it.

For those people who fly, I can understand that it's very important, indeed...

 

Anders

[+fizzymagic]

quote:

---

Originally posted by Kerry:

 

But really once manufacturers start stating accuracy figures around 3m and below then unless they also provide precision in the output to match (like at least 4 decimal minutes) there just blowing hot air.

---

 

Have you looked at the NMEA string coming out of your unit lately?

 

Didn't think so.

 

Kerry, you seem to be missing the point of this discussion. It is not "what is the absolute accuracy of a handlheld GPS unit?" but rather "does the use of WAAS improve the accuracy of handheld GPS units?"

 

The answer to the former question, at least for geocaching, is "no better than about 3 m." For more controlled situations, better accuracy can be achieved, but that accuracy is not really relevant to geocaching.

 

However, the latter question is what is of interest to geocachers. There has been a great deal of nonsense posted in these forums about how some believe that the use of WAAS actually decreases accuracy because it sometimes leads to a larger "accuracy" reading on Garmin units. As you have pointed out, the "accuracy" given by these units bears little resemblance to reality. The only way to test whether WAAS improves accuracy or not is to measure a known position with WAAS on and off.

 

Which is exactly what the person who started this thread did. He wanted to see if WAAS improved accuracy, so he measured it. As is the case with every test I have seen so far, WAAS did improve the accuracy significantly, despite the "accuracy" reading of the GPS.

 

So your claim that WAAS is useless on handlheld devices because they don't have centimeter accuracy is demonstrably incorrect.

[+azmark]

My GPSr is broke. It's display will not update. But on one point on earth it is 100% accruate 100% of the time.

[+azmark]

My GPSr is broke. It's display will not update. But on one point on earth it is 100% accruate 100% of the time.

[fig]

I guess my question has to do with accuracy, so Waas might apply here. I have read all the above post, and yes they are thouroughly fascinating..

 

My question is, if the accuracy is x number of feet, or meters, how in the world does my unit know when I am moving at .5 mph. I mean when I swing my arm holding my gps, it knows. Now I would think, without putting much brainpower to it, that if it was only accurate to a number of feet, that you would have to move that number of feet to be able to tell that there was movement. How does it have this incredible granularity for detecting movement, yet a much larger (proportionally) margin of error when it comes to pinpointing the exact location. Is it just all of the variables involved, the ones that Waas is trying to compensate for? I am just amazed at how the gps unit I have can sense small movements like extending my arm, yet not be that accurate when it comes to pinpointing your exact location.

 

Are GPS units that accurate as to granularity, but only accurate to exact location based on outside influences, for instance, fluctuations based on atmosphere? Am I making any sense here? In other words, when it says accurate to 12 feet, does that mean that the display coordinates can be + or - 12 feet from your actual location, and not that the unit can only tell where you are within 12 feet? I guess I am trying to determine if there are two different interpretations of accuracy here.

Fig (if you think you are confused...)

[+Alan2]

quote:

---

Originally posted by azmark:

My GPSr is broke. It's display will not update. But on one point on earth it is 100% accruate 100% of the time.

---

 

I left my broken watch in that spot. And when you find it, you will see that it is 100.00% accurate two times a day!

 

Alan

[+Alan2]

quote:

---

Originally posted by azmark:

My GPSr is broke. It's display will not update. But on one point on earth it is 100% accruate 100% of the time.

---

 

I left my broken watch in that spot. And when you find it, you will see that it is 100.00% accurate two times a day!

 

Alan

[+EraSeek]

quote:

---

Originally posted by JDMC:

And why hasn't anyone mentioned altitude? It is a 3 dimentional problem the GPSr has to solve, isn't it?

---

 

I have seen a description on this before but don't have it in front of me, but your unit will be less accurate in determining altitude then your hoizontal position. If I have this right, think of the GPS position accuracy as a kind of cone. Where the cone intersects with the ground, the accuracy is like a circle. That circle will be much smaller than the heighth of the cone at your position. So if you have an accuracy with a WAAS signal of 3 meters, the heighth of the cone will be something like 80 feet or 24 meters. These numbers are probably off somewhat. I think the manual says something like 49' horizontal 90 percent of the time (no WAAS) and 100' verticle.

 

I have a barometeric altimeter and frankly I think my GPS does just as good a job since SA was turned off. With the fluctations in barometric pressure, the error in my altimeter is often no less than my GPS.

 

[This message was edited by EraSeek on July 31, 2002 at 09:41 PM.]

[+EraSeek]

Kerry, what is SIS and CEP? I think you have mentioned before that averaging is a myth since SA was turned off, and I tend to agree, but you mentioned the channel myth. Explain if you would.

 

And FIG, you got me, maybe the big boys know.

[Kerry.]

EraSeek, gotta look for some data for Fizzymagic but these are easy ones.

 

SIS (Signal-In-Space) which accuracy is now based on, basically without the added "affects" that the user is responsible for managing. The system operators only control the signal at the satellite level and really have no control of some of the uncontrollables like atmospherics and yes "users".

 

CEP (Circular Error Probable) the radius of a circle within which contains 50% of the horizontal positions. SEP is (Spherical Error Probable) which is similar but is the radius of a sphere which relate to 50% of 3D positions.

 

The channel myth is that a 12 channel receiver is more accurate than a 10 or 9 or 8 or 6 or 5 etc considering things like satellite geometry etc. That is not so as really if receivers were configurable (or configured correctly) then one would mask out the low horizon satellites, which are the ones that can tend to decrease the accuracy. Some manufacturers actually do this as a fixed setting but others don't and will accept all satellites (regardless) right down to the horizon. There is a point to useable channels but really to the sales/marketing people more sounds better. Accuracy has little to do with the number of satellites (hence channels) and all to do with geometry. Really the full 12 channels are used only a small % of the time but right now we have an overpopulated system (28) so we are a little spoilt. WAAS is different but the 1 or 2 channels that WAAS uses is really irrelevent to the result. Similar with DGPS as there are only corrections transmitted for 9 satellites and the same satellites must be in view at both locations for corrections to be applied.

 

But to cut a long story short the "original" design was for a 11+1 but somebody got things wrong so everybody followed. Now would you have brought a 11+1 channel receiver if the opposition made a 12

 

Averaging certainly does very little for me these days.

 

Cheers, Kerry.

 

I never get lost everybody keeps telling me where to go

[Kerry.]

EraSeek, gotta look for some data for Fizzymagic but these are easy ones.

 

SIS (Signal-In-Space) which accuracy is now based on, basically without the added "affects" that the user is responsible for managing. The system operators only control the signal at the satellite level and really have no control of some of the uncontrollables like atmospherics and yes "users".

 

CEP (Circular Error Probable) the radius of a circle within which contains 50% of the horizontal positions. SEP is (Spherical Error Probable) which is similar but is the radius of a sphere which relate to 50% of 3D positions.

 

The channel myth is that a 12 channel receiver is more accurate than a 10 or 9 or 8 or 6 or 5 etc considering things like satellite geometry etc. That is not so as really if receivers were configurable (or configured correctly) then one would mask out the low horizon satellites, which are the ones that can tend to decrease the accuracy. Some manufacturers actually do this as a fixed setting but others don't and will accept all satellites (regardless) right down to the horizon. There is a point to useable channels but really to the sales/marketing people more sounds better. Accuracy has little to do with the number of satellites (hence channels) and all to do with geometry. Really the full 12 channels are used only a small % of the time but right now we have an overpopulated system (28) so we are a little spoilt. WAAS is different but the 1 or 2 channels that WAAS uses is really irrelevent to the result. Similar with DGPS as there are only corrections transmitted for 9 satellites and the same satellites must be in view at both locations for corrections to be applied.

 

But to cut a long story short the "original" design was for a 11+1 but somebody got things wrong so everybody followed. Now would you have brought a 11+1 channel receiver if the opposition made a 12

 

Averaging certainly does very little for me these days.

 

Cheers, Kerry.

 

I never get lost everybody keeps telling me where to go

[+apersson850]

Reading the NMEA output isn't easy on my Vista, provided that I don't bring a laptop or something else in the woods. Since it's too clumsy, I don't.

 

But you are right, fizzymagic, the additional decimal is there. Also, looking at the NMEA output, you can get the DOP value, not only CEP or whatever it is the unit is telling us.

 

It seems to me that at Garmin, they consider such information to be too complicated for the user to handle, when it comes to products like the eTrex line. Maybe they wanted these to be as simple as possible, well, simpler than the most possible complex, at least. Or people would consider all information shown as sheer magic (pun intended!).

 

Regarding positional vs. speed accuracy, the manufacturers seem to use a combination of methods to determine speed. The obvious method is to see where we were, and where we are, and how long time it took from one point to another. The other possibility is to measure the phase shift (Doppler effect) of the carrier signal, and measure the speed like that.

I haven't been able to sort out exactly how they do. To me, it seems that a Vista do use both methods, in different circumstances, but that's something that's virtually impossible to detect by just looking at the outside of the unit.

 

Still, these two different methods may account for how it can tell the speed when you wave your arm, and yet be uncertain about where your whole body is.

 

Since I can't see any effect of any augmentation on my home turf yet, I'll not comment on that.

 

By the way, on my way to work today, it showed the EPE as 131 meters (430 feet). And you say they are optimistic?

 

Anders

[+apersson850]

Reading the NMEA output isn't easy on my Vista, provided that I don't bring a laptop or something else in the woods. Since it's too clumsy, I don't.

 

But you are right, fizzymagic, the additional decimal is there. Also, looking at the NMEA output, you can get the DOP value, not only CEP or whatever it is the unit is telling us.

 

It seems to me that at Garmin, they consider such information to be too complicated for the user to handle, when it comes to products like the eTrex line. Maybe they wanted these to be as simple as possible, well, simpler than the most possible complex, at least. Or people would consider all information shown as sheer magic (pun intended!).

 

Regarding positional vs. speed accuracy, the manufacturers seem to use a combination of methods to determine speed. The obvious method is to see where we were, and where we are, and how long time it took from one point to another. The other possibility is to measure the phase shift (Doppler effect) of the carrier signal, and measure the speed like that.

I haven't been able to sort out exactly how they do. To me, it seems that a Vista do use both methods, in different circumstances, but that's something that's virtually impossible to detect by just looking at the outside of the unit.

 

Still, these two different methods may account for how it can tell the speed when you wave your arm, and yet be uncertain about where your whole body is.

 

Since I can't see any effect of any augmentation on my home turf yet, I'll not comment on that.

 

By the way, on my way to work today, it showed the EPE as 131 meters (430 feet). And you say they are optimistic?

 

Anders

[Kerry.]

quote:

---

Originally posted by fizzymagic:

 

So your claim that WAAS is useless on handlheld devices because they don't have centimeter accuracy is demonstrably incorrect.

---

 

No WAAS isn't all that it's made out to be especially in an environment it was never really intended and it has nothing to do with cm accuracy. One thing for sure these days there is no where near a US$6 billion improvment in accuracy. For a lot less than that amount of $$'s some more current generation sats and our coming second cival frequency will give more to many more users. The problem with getting new generation sats is the old ones appear to have been built too well and won't die quick enough

 

Anyway getting off the subject

 

The view that WAAS was never really intended for land based activities with obstructive environments is (now) basically universally accepted. To reduce the obstruction effects on especially the L1 WAAS frequency relying on a handheld with something like a patch antenna is at times expecting a little too much considering the requirements, constraints and possible conditions. I also wouldn’t limit WAAS’s distractions to simply handhelds either but handhelds have more limitations than some fixed mount units.

 

It’s only in recent times that much of the WAAS hype has been sorted out a little better and things better clarified. However there’s really more important things to WAAS than simply accuracy but recreational users really aren’t interested in those features and besides the receivers generally don’t have the software capability to optimize these benefits.

 

The following was done (deliberately) outside the ground station network for several reasons. 1) just to see what the "effects" would be and 2) to compare actual error (in the real world) to estimated error (EPE etc) for both standard GPS and the "affects" of just the WAAS carrier.

 

In both instances both the same receiver (recreational type) recorded over 23 hours on a well defined reference point.

 

Standard (real world) GPS accuracy was 1.8m (CEP), 2.4m (RMS) & 4m @ 95%. The receivers estimated position in comparison did not show less than 3.5m with 4.4m (CEP) and 5.0m 95% of the time. The estimated error was certainly very conservative but didn’t really reflect the real world. Strange the lowest value was 3.5m?

 

Based on the WAAS "received signal" the "affected" accuracy was 20.1m (CEP), 27.6m (RMS) and 87.3m @ 95%. The unit’s estimated errors were 100.5m (CEP), 133.7m (RMS) and simply heaps at 95% (max was just over 1km) which didn’t even come close to the real world as far as the unit knowing where it was or even thought it was. Obviously this isn't true WAAS accuarcy (as such) but does highlight was can happen when things go of the rails.

 

Visually, reality looked something like this

 

everybody keeps telling me where to go

[Kerry.]

quote:

---

Originally posted by fizzymagic:

 

So your claim that WAAS is useless on handlheld devices because they don't have centimeter accuracy is demonstrably incorrect.

---

 

No WAAS isn't all that it's made out to be especially in an environment it was never really intended and it has nothing to do with cm accuracy. One thing for sure these days there is no where near a US$6 billion improvment in accuracy. For a lot less than that amount of $$'s some more current generation sats and our coming second cival frequency will give more to many more users. The problem with getting new generation sats is the old ones appear to have been built too well and won't die quick enough

 

Anyway getting off the subject

 

The view that WAAS was never really intended for land based activities with obstructive environments is (now) basically universally accepted. To reduce the obstruction effects on especially the L1 WAAS frequency relying on a handheld with something like a patch antenna is at times expecting a little too much considering the requirements, constraints and possible conditions. I also wouldn’t limit WAAS’s distractions to simply handhelds either but handhelds have more limitations than some fixed mount units.

 

It’s only in recent times that much of the WAAS hype has been sorted out a little better and things better clarified. However there’s really more important things to WAAS than simply accuracy but recreational users really aren’t interested in those features and besides the receivers generally don’t have the software capability to optimize these benefits.

 

The following was done (deliberately) outside the ground station network for several reasons. 1) just to see what the "effects" would be and 2) to compare actual error (in the real world) to estimated error (EPE etc) for both standard GPS and the "affects" of just the WAAS carrier.

 

In both instances both the same receiver (recreational type) recorded over 23 hours on a well defined reference point.

 

Standard (real world) GPS accuracy was 1.8m (CEP), 2.4m (RMS) & 4m @ 95%. The receivers estimated position in comparison did not show less than 3.5m with 4.4m (CEP) and 5.0m 95% of the time. The estimated error was certainly very conservative but didn’t really reflect the real world. Strange the lowest value was 3.5m?

 

Based on the WAAS "received signal" the "affected" accuracy was 20.1m (CEP), 27.6m (RMS) and 87.3m @ 95%. The unit’s estimated errors were 100.5m (CEP), 133.7m (RMS) and simply heaps at 95% (max was just over 1km) which didn’t even come close to the real world as far as the unit knowing where it was or even thought it was. Obviously this isn't true WAAS accuarcy (as such) but does highlight was can happen when things go of the rails.

 

Visually, reality looked something like this

 

everybody keeps telling me where to go

[+BletchleyPark]

I have completely forgotten about the elevation testing that was done in the test described at the beginning of the thread!

 

As you may remember, the Benchmark's vertical position is listed as 938 feet. I first looked at the barometric reading on the Vista and it was exactly 938 ft! The amazing thing is that the last time I calibrated the elevation I did it at an general aviation airport control tower...about three months ago. Go figure! Perhaps, by coincidence, each day had very close to identical barometric pressure readings?

 

The last part of each test (before toggling WAAS on and off) was to read the GPS elevation from the satellite screen. Results follow:

 

Test 1 - no WAAS - GPS elevation: 952ft (minimum "bounce")

Test 2 - WAAS - GPS elevation: 957 ft (minimum bounce)

Test 3 - no WAAS - GPS elevation: ~944ft (some bounce between 941ft and 947ft)

Test 4 - WAAS - GPS elevation: ???ft (extreme bounce between 939ft and 953ft; very little stability)

 

I'm not sure why the "stability" of the elevation number degraded so much during the test. It's also interesting that the test producing the highest Lat/Lon positional accuracy (Test 2) yielded an elevation figure that was off by almost 20ft (one contour "interval" on Mapsource Topo maps). I'll take my elevation figure from the barometic pressure, thank you!

 

FWIW,

 

BP

[+BletchleyPark]

I have completely forgotten about the elevation testing that was done in the test described at the beginning of the thread!

 

As you may remember, the Benchmark's vertical position is listed as 938 feet. I first looked at the barometric reading on the Vista and it was exactly 938 ft! The amazing thing is that the last time I calibrated the elevation I did it at an general aviation airport control tower...about three months ago. Go figure! Perhaps, by coincidence, each day had very close to identical barometric pressure readings?

 

The last part of each test (before toggling WAAS on and off) was to read the GPS elevation from the satellite screen. Results follow:

 

Test 1 - no WAAS - GPS elevation: 952ft (minimum "bounce")

Test 2 - WAAS - GPS elevation: 957 ft (minimum bounce)

Test 3 - no WAAS - GPS elevation: ~944ft (some bounce between 941ft and 947ft)

Test 4 - WAAS - GPS elevation: ???ft (extreme bounce between 939ft and 953ft; very little stability)

 

I'm not sure why the "stability" of the elevation number degraded so much during the test. It's also interesting that the test producing the highest Lat/Lon positional accuracy (Test 2) yielded an elevation figure that was off by almost 20ft (one contour "interval" on Mapsource Topo maps). I'll take my elevation figure from the barometic pressure, thank you!

 

FWIW,

 

BP

[+fizzymagic]

quote:

---

Originally posted by Kerry:

 

The following was done (deliberately) outside the ground station network for several reasons. 1) just to see what the "effects" would be and 2) to compare actual error (in the real world) to estimated error (EPE etc) for both standard GPS and the "affects" of just the WAAS carrier.

 

---

 

In other words, you intentionally did a test of WAAS that was invalid. And you wanted to prove what?

 

Geez.

 

When I said that every test I knew of using WAAS showed improved accuracy, I assumed that people would understand that I meant every valid test.

 

Your test is the intellectual equivalent of claiming a Ferrari is a worthless car because you tested in on a 4WD-only road and it got stuck.

[+Prime Suspect]

quote:

---

Originally posted by BletchleyPark:

Yesterday, I performed an admittedly unscientific test of the WAAS feature of my Vista. The results were interesting and may indicate that the traditional "Accuracy" (in feet) figure is not necessarily the only measure of the WAAS advantage.

---

Nowhere in your original message do I see any mention of how many satellites you had collected WAAS data on. On the few occasions where I've gotten WAAS data on all satellites I have a lock on, my "accuracy" reading is typically in the 5' to 10' range. Your double-digit readings lead me to believe that you never collected WAAS data on all the usable satellites.