Different or Same???

[+trippy1976]

Kerry,

 

Does your graph (and post) indicate that 95% of the time, the GPSr will be within 31' (ish)?

 

I think that's a pretty impressive number in and of itself. I consider 20 feet to be 'a few feet' and really... in the scheme of world-wide navigation - that's pretty good IMHO Hehehe. Columbus wishes he was so lucky!

 

Close enough to get to a geocache.

 

I had wondered based on our little experiment whether satellite positioning played a part in readings that vary according to the time. Because in our experiment Rusty did his reading around 8:30am and the rest of us did it at 11:30, then Rusty re-measured and found that the point had 'drifted' about 35 feet. I see in your graph some pretty dramatic changes in a relatively short period that would seem to lend support to this theory. I've also gone to geocaches where the logs by other people were 'coordinates took me right there' while my attempt was done at night and I was much further off (sometimes up to 100 feet off).

 

Do you think this could be the case? Are the GPS satellites in a synchronous orbit or no? Could be all hogwash, but this has been my observation.

 

--------

trippy1976 - Team KKF2A

[+GOT GPS?]

I have found that one GPS V might be worst than another, or one Vista would be worst than another.

Best way to test them, would be to put the same type of battery in each GPS, then see how each unit can recieve satellite info inside a house near a window. Another thing is to try different brands of batteries in the gps and do a shake-the-gps test(to see if they shut off with different brands).

 

The GPS units would have to be all tested with the same satellite positions in the sky, and preferably be 4 sats overhead, with others scattered about in the sky.

 

----------------------------

 

I have found that Energizer Alkalines are the best battery, because they are slightly larger than others, so they fit snuggly in the compartment of the gps.

 

I have found also that the processor may be weak in some GPS V's, like mine. It has an excelent antenna, but is more quirky than most units with it's processor. could be because it has hit the ground a couple of times.

 

Age of the electronics is a factor also. Old units can be sluggish.

 

I have compared units side by side, and find it hard to do, without a professonal GPS unit.

 

Only GOD knows whats best: A trimble GPS.

 

Factors involved in poor GPS expirience:

*Antenna wire too long

 

*Antenna connector dirty

 

*Using Batteries that dont fit snuggly. They will rattle when you shake the gps.

 

*Unit with a bad software version.

 

*Allowing the tracklog to completely fill, causing possible errors.

 

*Having too many features inabled like the data port being left on, when not needed(eats batteries)

 

*Trying to use the GPS when there are almost no satellites directly overhead. Wish the GPS could give you the best times that it will work best. This would be better than the Best times to Fish Predictor.

 

*Not knowing enough about GPS units. Pick up a basic guide on GPS technology, on how to get the most out of the GPS.

 

*Not studying a downloadable User Manual before buying that GPS that you are wanting to buy.

 

Like my 2000 page book on this subject

 

------------------------

My home page about GPS units and information

[+Mopar]

quote:

---

Originally posted by Alan2:

Let me give you another analogy. You didn’t like my last one about weight and scales. Let’s say I’m target shooting. You have a motion detector tracking my aim point. If you look at the graph it produces, I’ll be all over the target. You than draw the conclusion there’s no way I can consistently hit the bulls-eye. The graph “proves” it. So I shoot off ten rounds and amazingly 7 are in the bulls-eye, two are in the next ring and one is off target. You say that can’t be. But I waited until I knew I was on target before squeezing off the shot.

 

Same thing with the GPS. You’re looking at the graph over a period of time. You’re not considering human intelligence or action in this case waiting for the signals to “settle” just like my squeeze point with the rifle.

---

Hey, that's a great analogy, but it has one flaw. The GPS doesnt have any way of knowing when its "pointed at the bullseye". It has no way of knowing where it is except for data that may or may not be accurate. So to make your analogy more accurate, you need to wear a blindfold while you are aiming your rifle. While you are shooting someone standing next to you can tell you how far you MIGHT be from the bullseye. Now lets see that target.

 

Illegitimus non carborundum!

[Kerry.]

Trippy1976, I think you've realised that time of day matters and it does as the satellites are in roughly 12 hour geocentric orbits and yes between 8.30am and 11.30 things will certainly have changed, sometimes for the better sometimes for the worse but always generally with the system specifications (95% of the time of course). The following link attempts to outline exactly your thinking/observation

 

http://www.cqnet.com.au/~user/mattk/gps/gps_plan.htm

 

Basically if nothing changes in the system then things should be "similar" from one day to the next however this doesn't take into account other things (like atmospherics etc) that also can change over time. Synchronous orbit? no not really as even if the system isn't touched things will change from day to day.

 

However on most days maintainence is carried out on individual satellites so these are unavailable, which for that period they are unavailable the geometry, accuracy and other things are affected. Sometimes these shutdown warnings are made in advance and can be planned around where as other times one doesn't know a satellite is/was bad until after the event. If the receiever didn't pick up this problem sat (and most stand-alone units can't) then that problem gets used in the position solution.

 

The 95% accuracy of the last graph (Doy 1153 June 2, 2001) was actually 19 feet, which isn't bad at all. The one prior to that (Oct 4/5) was 25 feet and 31 feet, which is still well inside the system spec of 42' @95% SIS.

 

As for 20' well that's a bit more than a few feet (actually a bit more than a few metres) but this general discussion on accuracy focused on the claim (assumption) of 6-8 feet "all the time", "everywhere", which is certainly a myth.

 

Cheers, Kerry.

 

I never get lost everybody keeps telling me where to go

[+Alan2]

Mopar: I believe the GPS is not blindfolded.

 

It does have a pretty reliable range just how accurate it is going to be within 95% of the time. Anders mentioned DOP. I'll leave it it to him to explain DOP because I frankly don't know what it is.

 

But I'm sure it makes sense if you look at it another way. IF the GPS is reading 8 strong signals, it "knows" or the programmer from past experience knows, that there's lets say a 95% chance the circle of error is 18 feet. Whereas if its reading 5 sats at half strength, the circle should be 85 feet. And that's what the unit displays. Don't you use the error info when you're out geocaching? I do -all the time. How else can I know if the coordinates I'm reading are way, way off and not reliable. So I wait until the error comes down to it's minimum and then track from that point. Those markers given by the gps mfr are not just to sell units but actually provide a vary good guideline when the GPS is reporting acccurate readings.

 

I believe geocachers who are constantly using GPS's get a tremendous feel when their GPS's are on target. They have experience and watching the GPS as they approach the cache and after a while, they can look at the error indicator, look at the coordinates, and know at what moment they can mark an very accurate reading. And this is why trippy's friends and the 12 people at my benchmark cache got readings so close to actual.

 

All wagering aside, I think it would be very interesing and enlgihtening to test this hypothesis further. Isn't anyone beside me curious about this? Wouldn't it be interesting and important to know just how accurate our equipment is in the field? Would someone else like to set up a benchmark virtual similar to mine that can test the theory?

 

Alan

[+Alan2]

quote:

---

Alan, actually it's the night after the (your) morning before so your just getting into Thursday and here Thursday has just about gone.

---

 

Now I know why we were at such odds. It seems we're on different planets. I really appreciated the spirited discussion we had.

By the way, I'll send my money to my favorite charity but I was hoping to include your bet too.

 

PLease help me explain DOP per the previous post and other things that help the GPS to know how accurate it can determine it is at one time or another. And I'm still looking for you to contribute to our future experiments - if your game. No wagering of course.

 

Alan

[Kerry.]

Alan, different planets yeah some see it like that but we're all using the same "global" system and the world's very first true global utility .

 

As for the DOP check at the following post (I've yet to track down what I'm looking for) and hopefully the software will be useful in helping this discussion, especially with regards the relationship between DOP, satellites and accuracy (to a point).

 

One must remember that simply having X number of satellites in view doesn't dictate what the accuracy will be (totally).

 

Hopefully this software will help compare these things realizing that the geometry of the satellites fixes the DOP value and that DOP value has a major influence (but not all influence) on the expected accuracy of the position.

 

Cheers, Kerry.

 

I never get lost everybody keeps telling me where to go

[+Mopar]

quote:

---

Originally posted by Alan2:

Mopar: I believe the GPS is not blindfolded.

 

It does have a pretty reliable range just how accurate it is going to be within 95% of the time. Anders mentioned DOP. I'll leave it it to him to explain DOP because I frankly don't know what it is.

 

But I'm sure it makes sense if you look at it another way. IF the GPS is reading 8 strong signals, it "knows" or the programmer from past experience knows, that there's lets say a 95% chance the circle of error is 18 feet. Whereas if its reading 5 sats at half strength, the circle should be 85 feet. And that's what the unit displays. Don't you use the error info when you're out geocaching? I do -all the time. How else can I know if the coordinates I'm reading are way, way off and not reliable. So I wait until the error comes down to it's minimum and then track from that point. Those markers given by the gps mfr are not just to sell units but actually provide a vary good guideline when the GPS is reporting acccurate readings.

 

I believe geocachers who are constantly using GPS's get a tremendous feel when their GPS's are on target. They have experience and watching the GPS as they approach the cache and after a while, they can look at the error indicator, look at the coordinates, and know at what moment they can mark an very accurate reading. And this is why trippy's friends and the 12 people at my benchmark cache got readings so close to actual.

 

All wagering aside, I think it would be very interesing and enlgihtening to test this hypothesis further. Isn't anyone beside me curious about this? Wouldn't it be interesting and important to know just how accurate our equipment is in the field? Would someone else like to set up a benchmark virtual similar to mine that can test the theory?

 

Alan

---

 

Ah, but you nailed the problem in your second sentence.

quote:

---

It does have a pretty reliable range just how accurate it is going to be within 95% of the time.

---

The other 5% is the problem. That means hat 5% of the time, even though your GPS THINKS its accurate (low EPE), it's really not. Sure, if you are atanding at a known location (a benchmark) this isnt a problem, because you know the GPS is reading wrong, and dismiss that reading and wait until you get a reading that agrees with the benchmark. Without that benchmark, you have know way of knowing if you are in the 95% section or the 5% section at any one particular time. If you stand in that one spot, and avg the readings, then eventually you will factor out that 5% error.

It was suggested some posts back that you leave your GPS stationary and record a tracklog. Try it. Its actually pretty amazing. Do it in a clear area where you have consistant "accurate" readings. after 15-20 minutes the tracklog will still wander all over the area.

You mention real-world tests. I have a couple of real-world examples for you.

Havent you ever gone for a cache, had a great signal, followed the pointer right to ground zero, and the cache wasn't there? Of course. I've stood there with my GPS showing 9ft of accuracy, and no cache in site. Look at GPS again, and its now showing the cache 60ft away, and 11ft of accuracy. Which reading is correct?

2nd example: friend and I are out caching together. Garmin 76 and V. Ecellent coverage, usually 7-15ft error. We were consistently 50-75 feet off from each other the whole day. When it came to actually finding the cache, sometimes it was mine that was close, sometimes his.

3rd example. Another day caching, this time garmin 76 and a Meridian. Poor overage, accuracy showing 25-50ft most times. GPS's were usually 150ft and often more then 300ft off from each other! It was actually enough to make us stop and check that one or the other hadnt got swithed to a different datum somehow. Neither GPS read closer then 50ft to the actual caches we found that day.

4th example. I placed a cache last winter. It was close to home, so it was easy to take multiple readings before posting it. I left the GPS sit and avg 500 readings, then shut it off, let it reaquire and avg 500 more readings. Came back and did this 2 more days. Then avged the 6 averaged readings together. Total of 3000 readings avged. I was pretty dadgum proud of myself, and most finders comment on how close their GPS gets them. MOST. There is still a handful that complains just the opposite, that try as they might, my readings were way off.

I do alot of caching in a coastal area. I often see 5-7ft accuracy, and have a solid WAAS lock. You would think that I should be able to walk right up to every cache I hunt if the GPS is really accurate to 5 ft. Usually, I get about 50ft away, shoot a bearing with the compass, then move to another location and do the same. Where they cross i drop my pack and the GPS, and start working my way out from there. After the 1st 15-20 caches of lots of wandering around staring at the GPS, I learned to stop worshipping that dadgum pointer and just look.

Oh, and a fairly good description of DOP can be found at:

http://joe.mehaffey.com/dopnontech.htm

 

Illegitimus non carborundum!

[Kerry.]

Mission Planning Software

 

Really not entirely relevant to the type of equipment or purpose used here but this type of software might help in understanding the relationship between number of satellites, satellite geometry, DOP and to some extent accuracy and a few other things.

 

Probably useful for knowing when not to be at a specific location more so than when one should be there as minimal conditions are generally in the minority these days but probably the most important times to avoid if one wants reliable confident results. In other words when the number of Sats are low and the DOP goes of the top of the graph one really doesn’t want to be expecting good results. Actually there’s more chance of reasonable results when the DOP is low regardless of the number of Sats.

 

Click on the following URL and under Mission Planning click on MP_WIN, which should go to a FTP site and then down load the file MP_WIN.ZIP which is the main software (compressed).

 

http://www.thales-geosolutions.com/skyfix/gpsinfo.html

 

As well as this software one will need to regularly maintain an updated almanac file which can also be downloaded from the same page. Just note the link on this page appears to have an error so use the following FTP link and copy the ALMYY.DOY file (YY=year and this year will be 02, DOY will be something like .290 or whatever the Day Of Year was the almanac was created on) to the directory where the software is installed. If there’s been no system changes then predictions can be made into the immediate future without problems if the almanac is reasonably up to date.

 

ftp://ftp.ashtech.com/almanacs/

 

The software contains a default list of location files but users can create their own specific locations, which only have to be approximate anyway.

 

Specific settings which will affect the output is masking angle (not all receivers have the same settings) and if necessary obstructions can be added which can change the whole scenario. Setting a mask angle around 5 (or 10 degrees) is best anyway as that probably gives a better indication of real (minimum) conditions anyway.

 

Cheers, Kerry.

 

I never get lost everybody keeps telling me where to go

[+Alan2]

Mopar: You make some really good points. The idea that the person would wait for the GPS to match the known benchmark position or get as close as possible before he marked it would distort the results. As I recall, I didn't do that - I just left my Vista on the benchmark for a minute or so and then marked the waypoint a number of times and reported those results which were within feet. However I don't know what the other 11 people did. It would b interesting to see if I repeated what I did on a number of different days to see if the results change. I'll try doing that and get back with the data.

 

Regarding actual caching experiences, your right about readings bouncing all over the place at actual conventional caches. But since most the ones I've hunted have been in cover and all had the locations marked by other GPS's meaning that the ccoordinates are assumes "wrong" because they weren't benchmarks, I could have overly discounted the coordinates bouncing. This is why I can only see us doing a real world test in clear sky areas as we know that signal distortion in covred areas do make the results consistantly inaccurate.

 

One question though. Are you saying though that in clear cover with a GPS stable, you will get readings within let's say fifteen feet 95% of the time? If that's true then that's pretty good odds although I argree with you that not knowing when the reading is in the 5% could be a bummer. Being right 95% of the time would be fantastic at the race track but could cause problems if your life depended on 100% of the time.

 

We get back to the definition of the word "consistancy".

 

Alan

[+Alan2]

One clarification to my last post. While I didn't compare the coordinates I was getting to benchmark actual before marking the waypoints in my GPS, I did check first to see that the circle of error was at its minimum. I believe it was around 18-19 feet except for the one Wass when the error was 16 feet. And I tried not blocking the GPS with my body when I hit the mark button.

 

Alan

[+apersson850]

Several people are referring to the signal strength vs. the EPE. Why? As far as I can understand, that has nothing to do with it. The pseudocode matching is analog, which means that the matching is easier if the signal strength is higher. But as long as there hasn't been any matching, the signal isn't used in the PVT solution. On the other hand, if the unit managed to lock onto the satellite, then the signal strength shouldn't make much of a difference, since it will still be within the resolution of the pseudocode. Which means that the information is digital, and doesn't deteoriate with lower signal strength. No handheld GPS uses the carrier phase detection anyway.

 

So, while signal strength certainly affects how many satellites can be received, and thus indirectly the DOP and therefore the EPE, I can't see that signal strength would be of importance for satellites that actually are used for position calculation, even if their signal strength bars hardly rise above the baseline they are drawn upon.

 

Or am I really barking up the wrong satellite here?

 

Anders

[+Alan2]

Then why do the sats with the biggest signal strength seem to lock "black" first?

 

Alan

[Kerry.]

As I understand it signal strength is used in conjunction with other data like elevation, DOP etc to "decide" just what sats it "should" best use if/when there's a choice.

 

Basically it should use the best geometry from the strongest satellites as the receiver generally doesn't know what's the cause of the weak signal (it can't see).

 

A 12 channel reciever probably doesn't rely on things as much as a say a 5 or 8 channel reciever but the underlying decision principle is probably much the same and also based on other data that's probably not as obvious as signal strength, DOP etc.

 

Lock appears to pick on the easy ones first, starts navigating then sorts out what to do with the rest in decreasing priority relative to configuration parameters (like mask angles), number of channels etc.

 

Cheers, Kerry.

 

I never get lost everybody keeps telling me where to go

[+apersson850]

quote:

---

Originally posted by Alan2:

Then why do the sats with the biggest signal strength seem to lock "black" first?

 

Alan

---

That's exactly what I said. They are easier to lock on to, but once the lock has been established signal strength doesn't matter, as far as I can understand. If it's locked, it's locked. Keeping the lock is of course also easier when the signal is stronger, but I still can't see any relation between the signal strength of a locked satellite and its usefulness.

 

If the receiver makes any kind of judgement regarding which channels to use, I don't know. But to me, it seems unlikely that a twelve channel receiver wouldn't use all channels, if that many satellites are above the horizon. The only decision that has to be made is if the almanac predicts more than twelve sats above the horizon. In that case I would say that omitting these very close to the horizon (or rather the mask anlge) makes most sense, since they'll probably not be too receivable anyway. The only objection to that is that one may think of a constellation, where including a very low sat would substantially contribute to the decreasing of the DOP, assuming that most other sats are grouped elsewhere on the sky. But that doesn't seem too likely either. Or perhaps it is, nearer the poles? But I don't want to go there, anyway...

 

Kerry's notice that lock starts with the easy ones, then goes outward in decreasing priority, I consider not an intentional strategy, but a consequential one. The "easy" ones are easier to lock on to, by definition. But I think the unit locks on to all available satellites as soon as it can sort out the pseudocode. Again, with a twelve channel parallel receiver, the only prioritizing that has to be done is which twelve to use, when more than that are available.

 

Perhaps one can figure out something about this strategy by observing how the unit omits satellites, when WAAS is enabled. If there already are twelve satellites receivable at that instance, then two of them has to go, to leave room for the augmenting sats.

 

Anders

[Kerry.]

quote:

---

Originally posted by Anders:

.... Perhaps one can figure out something about this strategy by observing how the unit omits satellites, when WAAS is enabled. If there already are twelve satellites receivable at that instance, then two of them has to go, to leave room for the augmenting sats.

---

 

With WAAS or Differential that's not a strategy decision as a receiver simply can't use a differential correction if the satellite is not common at BOTH the reference site and the receiver.

 

Differential (dGPS) only transmit corrections for the best 9 satellites if the receiver is in dGPS mode then the decision is easy, if it's not in view then it won't/can't be used.

 

Cheers, Kerry.

 

I never get lost everybody keeps telling me where to go

[+apersson850]

You're right, Kerry, that makes sense. Didn't think about that it had to be visible at two different places.

 

Anders

[Kerry.]

Anders, with regards mask angles some recent changes announced to the dGPS configuration, which undoubtedly will get adopted wordwide

 

http://www.navcen.uscg.gov/news/MaskAngle.htm

 

Cheers, Kerry.

 

I never get lost everybody keeps telling me where to go

[+apersson850]

To add to this mask angle thing, it has recently become obvious to me, that my GPS (eTrex Vista) uses different mask angles, depending upon if I use it in Normal or Battery save mode.

 

When running in Battery save mode, the mask angle is raised, thus the unit isn't bothering about trying to find sats very low on the horizon.

Probably the receiving circuits can be turned off individually, or it's simply a question about that they don't keep all twelve channels running for a longer time, as they otherwise would have been forced to do, to sort out the weak signals from low flying objects.

 

Anders

[paradox0430]

I have to say that as much as I love my Garmin V, the antenna leaves a little to be desired - loses signal often.

[+Brokenwing]

Interesting discussion, but getting back to the original conjecture that all GPSr's are the same, I wonder how many of you have seen this:

 

http://www.gpsinformation.net/mgoldreview/sportrak-pro-trees.html

 

http://www.gpsinformation.net/mgoldreview/sportrak-pro-survey.html

 

http://www.gpsinformation.net/mgoldreview/map-gold.html

 

The above test data from Jack Yeazel seem to suggest there are indeed differences in handheld GPS receivers.

 

Scott / Brokenwing

 

-There's a thin line between geocaching and walking in circles like an idiot.-