Kerry.
-
Posts
1125 -
Joined
-
Last visited
Posts posted by Kerry.
-
-
Yes Dave, one scenario of many and many is really an understatement. The Dynamic nature of GPS basically eliminates anything actually occuring exactly again.
So how about
5 4 3 2 3 4 5 call that 3.7, better than 5.
2 3 4 5 4 3 2 call that 3.3, worse than 2
One could come up with what ifs all day but one thing is for sure the track nature is not predictable but neither is it random.
-
For those interested in what the other 5% can/could/might look like then THIS IS WHAT CAN ACTUALLY OCCUR
This is really an extreme example but it is real and there's nothing stopping similar scenario's within what might be classed as difficult to detect at the time.
Cheers, Kerry.
-
Well, then that makes me very curious how anyone could have wrong coordinates. You take an eTrex Vista and "casually" mark a waypoint, yet there are caches 50-100 off. How's that?
The world average system spec is less than 43 feet 95% of the time, there another 5% that is unknown (more than 43 feet). The worst case scenario is less than 118 feet at 95% again there is an unknown 5% that "could" be more than 118 feet.
Accuracy, geometry and many conditions in general that affect accuracy are in fact quite variable and barring the "affect of the user" on accuracy all that is quite possible especially considering the system accuracy specs (43 and 118 feet etc) is actually Signal-In-Space accuracy before the user does what some users do.
There can be many many reasons and really in some cases there's an appliance thinking that is applied with GPS where some believe it's always perfect and one only has to point and push and it has to be right.
For a period of time a position could certainly be off by any number and there is some classic examples of quite large distances for quite long period of times and unfortuneately the unwary wouldn't pick this up and if one was averaging a group of position solutions that were 40km off then the end result is simply going to be 40km incorrect as well. There's no magic formula where a set of incorrect position solutions suddenly become correct.
Cheers, Kerry.
-
However, I think the crux is this. What is the likelyhood of any one reading being more accurate than several readings averaged together?
Maybe think about it this way. Over a period of time the position track will create a cob web affect and roughly the average of many readings (the minimum spec is 24 hours) will/should approximate close to the actual real world position, give or take a bit.
To create this cobweb effect the position solution really has to spend half the time going one way and half the time coming back.
Over a short period of time and at the rate of change of the position solution now that SA is discontinued it really doesn't go all that far in a short period of time. The position solution now that SA is discontinued is quite stable and one reason why differential stations don't really have to transmit corrections every second or two and could transmit a correction every 10 seconds etc has not a lot changes in this time frame compared to the required end result.
So if say the first position is say 5 metres north and the position solution tracks north (4cm/sec) for the next 60 seconds and your averaging then your position will technically be 6.2 metres off and worse then the first original position. Of course if it just happened to be tracking south then the position would be 3.8 metres and better.
There is absolutely no way that the position solution is by some magical method always tracking closer to the absolute position. Half the time is basically spent wandering out and the other half wandering back again and this is where short term averaging fails completely and has no integrity as to if the end result is better or worse. It's really a lucky dip.
The thing is you don't know which way the position is tracking relative to/or the absolute coordinates.
Cheers, Kerry.
-
The one thing that leaves averaging for dead these days is actually being in the required location at the BEST time. What's the point of going to all the trouble of going to a location just to have to work with crap geometry and expecting averaging is going to fix bad geometry, high PDOP etc.
Also take the current position rate of change now that Selective Availability is discontinued. This is something like 4cm per second or around 2.4 metres over 1 minute. So even if the position is tracking closer towards the absolute real world position (which one doesn't know anyway, either the absolute position or the track trend for better or for worse) then if one averages for 1 minute the maximum affect on position assuming a straight track can be no better than 1.2 metres and considering the display precision of just about every handheld then in reality it's a waste of time.
Cheers, Kerry.
-
Like I said if averaging feels good then do it and some can argue all day about that but without taking into full account the actual characterics of GPS behaviour then some of the thinking and comments about random behaviour is off base.
GPS behaviour is neither random or predictable with respect to Service Availability, Service Reliability and Accuracy, which must all be taken into account for the result to mean anything. The vast majority of the time the track is reasonably constant in rate but variable in direction and with SA discontinued over the period of time some recommend to average the end result simply can't be guaranteed as being better.
A GPS position solution simply does not jump all over the place but basically has a slow undertermined track and this is where many get caught up in the averaging myth and especially now that SA is discontinued.
Cheers, Kerry.
-
I should have known when I posted some valid technical information Kerry would find some way to post misinformation just to be contradictory.
He is almost completely wrong, folks. There is a portion of the GPS error that is not random, but it changes over time in such a way that it can be minimized. Averaging is a frequently-used tool to improve position accuracy. What's particularly amusing about Kerry's little statement above is that he has frequently pointed out in the past that extensive averaging must be used in order to obtain centimeter-accuracy GPS positions. So he's even contradicted himself.
I purposely didn't get into an even more complex discussion about systematic versus random errors and how they apply to GPS readings, because I wanted to give information that would actually be useful to people.
I can now see that I wasted a significant amount of time for nothing.
Forget it.
For s start Fizzy "extensive averaging" is really not what allows centimetre accuracy (as such) not in your context anyway. This is fact not a contradiction, just that one needs to understand the difference between the two.
But yes "extensive averaging" in your context/understanding of averaging does reduce possible errors. 95% of all possible eroors can be averaged out in 24 hours, 99.99% of all errors requires about 30 days. Simply not feasible from a preactical point of view is it.
Information that is actually useful to people is what actually occurs in the real world and really you've only wasted your time if you can't accept the facts based on actual data.
Basically there appears this thinking that one only has to average and presto the coordinates will be better. Sure they can be better BUT they can also can be worse and which way the pendulum swings for someone averaging is a toss and people have to realize this and not simply average for the sake of averaging, it doesn't always work but it does work about 50% of the time, the other 50% of the time it simply doesn't.
The point NavDog makes with obstructions and geometry is also valid and especially with obstructions there is almost infinite possibilities.
Affect of Obstructions on Averaging inn the real world
GPS averaging in general in the Real world
More on Averaging in the Real World
People need to decide for themselves if averaging today without Selective Availability, in obstructed environements etc is actually viable and all the technical theory in the world doesn't replace on the ground data. To simply average for the sake of averaging most people are simply fooling themselves as quite frankly without hindsight they wouldn't have a clue if the coordinates are better ot worse.
Cheers, Kerry.
-
Simple. Average several measurements. For those measurement errors that are random, averaging is very effective.
GPS errors are not random and averaging gives absolutely NO guarantee that the final positions will be more accurate.
Cheers, Kerry.
-
So if someone has zero (0) finds and posts what might be the relevance of that, what might that mean and really is it actually important at all
Cheers, Kerry.
-
3 satellites will not give a position in three dimensional space. 3 sats will give a two dimensional fix with the Z being assumed and if Z is assumed incorrect then the 2D fix isn't worth much either.
It needs 4 to compute all unknowns X, Y, Z and Time. 3 sats will generally solve for X, Y and Time.
Catch 22 with low horizon sats is that these introduce the most error due to atmospheric influences as the signal has to pass through more of the atmosphere the lower the satellite is to the horizon.
Ideally receivers should cut-off these low horizon sats or at least allow the user to define a cut-off angle.
Cheers, Kerry.
-
Aren't GPS sats. in a geosyncronous orbit? Meaning they don't revolve around the earth but stay in the same spot?
Oh no, they move alright, the system simply wouldn't work if they didn't orbit. They travel at something like 4000m/s and effectively orbit the earth twice a day.
As for the 1500mph, GPS isn't perfect, no measuring system is and what the other unknown 5% of the accuracy specifications tend to cover, the unexpected for whatever reason. Another reason why the specifications are for Signal In Space (SIS) conditions as one has no control how the user will/might implement/use things at their end.
Cheers, Kerry.
-
Sub-metre accuracy isn't limited to dual freq receivers as many single freq receivers can achieve sub-metre even cm accuracy just that it takes a bit more time.
Once accuracy (with integrity) towards the sub-metre, cm mark is required then irt will require some form of differential/static/RTK/post-processing of some sort and the limiting factors basically come down to distance and/or time. For real-time systems (and generally these will be dual freq) the the limiting factor is basically distance with modern technology/software now being able to resolve ambiguities out to about 30km.
Static/post-processing can resolve sub cm for hundreds of kilometres providing there has been enough data recorded and this comes down to time. Longer the distance the more the time required and sometimes this can be many days, even with dual freq.
In the above link http://products.thalesnavigation.com/en/pr...t.asp?PRODID=40 this is a single freq receiver spec'd to sub cm (< 0.01m) accuracy but is limited to 15km and to achieve this accuracy requires a minimum recording time of 10 minutes+1m per km so for 15km <1cm accuracy would require "at least" 25 minutes of recording time. A dual freq receiver would know over less than 1cm over this distance in much less time than that but would still require a minimum amount of recorded data to achieve this.
Single freq receiver/s can achieve sub cm accuracy but takes time and is obviously a cheaper option but if the job required positions from the middle of a busy road then having to setup for 25 minutes isn't practical at all.
Bons, with differential the principle is that both receivers are recording/using the same satellites at precisely the same time (both receivers running and recording at the same time), hence any inaccuracies (in simple terms) in the atmospherics etc is assumed to cancel out so one has a relative (correct) vector and by knowing the actual position of the abse then then position of the required point can be determined absed on relatively. The further the distance between the 2 receivers then the more possibility there is of differing atmospherics (plus a few other issues) so the longer the distance the more time/data is required with the longer distances the less suitable are single freq receivers.
Cheers, Kerry.
-
GPSIIR-11 has now been declared fully operational and useable.
Cheers, Kerry.
-
First some need to get past the misconception that military units are more accurate than civil units (as such). There days without Selective Availability there is very little difference between standard military and civil receivers.
Sub-metre accuracy of any receiver, be it a Trimble or any other has to be put into proper perspective. For sure with single freq receivers this can be achieved using augmentation or post processing of some sort. With dual freq using specific methods and techniques sub centimetre is quite easily achieved.
Accuracy is dependent on receiver type, software, antenna's and a lot of other things including technique and what they can with the information they receive and the information they receive is basically the same information that everybody receives (to a point relative to the receiver). When referring to accuracy there needs to be some definition/clarification of just what type of accuracy they are referring.
Basically you can have what you pay for and then generally need some understanding in knowing how to use it.
Cheers, Kerry.
-
.... Lat/Long are cartesian coordinates, that's something completely different
Yes, Cartesian coordinates are something completely different but no way are Lat/Long the same as cartesian coordinates.
Cheers, Kerry.
-
There can be several reasons but was the distance was shorter or longer.
Cheers, Kerry.
-
Also WAAS seems to have no effect on the accuracy.
I'm wondering if WAAS is at fault, atmospherics, the GPS or me at fault.
I live in Bristol, England. Anyone know a way of getting the general status of the GPS accuracy in a region? A government website maybe?
The equiv to WAAS in your part of the world is EGNOS and works exactly the same.
EGNOS is still only transmitting a test signal so could have its ups and downs.
General status of GPS accuracy can be obtained in hindsight, otherwise one must assume the system accuracy specs of less than 13 metres SIS, 95% of the time world average or worst case scenario 36 metres SIS 95% of the time.
Cheers, Kerry.
-
Is it possible for a GPS unit to experience a constant "drift" in one direction or another??
No, but many things cause GPS drift and all combine to account for an unpredictable error rate that is generally less than a specificed value but not predictable in direction.
Cheers, Kerry.
-
therotical accuracy Ok therotical accuracy is well defined in the system specifications and what you probably mean is "therotical precision".
Theoretically 1 second is around 31 metres (~ 101 feet), 0.1" is then roughly 3 metres and 0.01" around 0.3 metres but is there any recreational GPS that displays to 0.01"?
Many recreational GPS units these days will output via NMEA to 0.0001', as in minutes as NMEA doesn't support Deg/min/Sec output anyway. Theoretically that (0.0001') is around 0.2 metres but considering GPS SPS accuracy is nowhere even near that capable then yes it's really only theoretical.
Cheers, Kerry.
-
My question is will my Magellan Meridian know to read this new satellite (and if so, what number will it show up as on my GPS screen), or will I need a new software upgrade?
Once a new satellite is finalized and checked out (normally takes 4 weeks or more) it will be set to healthy and GPS receivers will automatically pick this fact up the next time it receives almanac data, which it is basically doing all the time anyway. Just do nothing and when it comes on line it will automatically show up.
It will show as (PRN) 19.
Cheers, Kerry.
-
Not sure what type of accuracy you were expecting but there's no way known you are going to calibrate a GPS that "appears 10 feet off". SPS GPS accuracy simply isn't that good and even if there was a calibration function (which basically there isn't) it would serve no purpose.
Cheers, Kerry.
-
1991 Lowrance GlobalNav.
Cheers, Kerry.
-
"GPS 2R-11, also known as Space Vehicle No. 59, is expected to complete post-launch testing and enter service by mid-April. It is destined to fill the Plane C, Slot 3 position in the GPS network.
The spacecraft being replaced -- GPS 2A-19 or Space Vehicle No. 31 -- was launched 11 years ago this month."
Let's see 31 goes into 59... (carry the 2)... 1.9032258 times... that's gonna leave a mark!
Yes, (now checked a little further) GPS 2A-19 is still sitting in C3 and is SVN 31 and it's also PRN 31, so GPS 2R-11 is SVN 59 but I don't recall seeing anything anywhere yet about its PRN number? Could become PRN31 but normally this isn't made available until just before being brought on line?
Actually 2A-19 is still showing in the almanac? I'm using (only a week or so old) but I thought with all the problems from 19 several months ago it had been shutdown as it really went wild in early Jan, 2004.
So obviously they have to kick 19 out (if in fact it hasn't already been done) in the next few weeks.
Cheers, Kerry.
-
Is there a logical explanation to these "errors" with my GPS?
There could/can be many reasons why this might occur, GPS is not guaranteed 100%
Can it have anything to do with the military possibly degrading the signals during the war?
What war? but in any case, No forget this possibility.
Cheers, Kerry.
Accuracy Of Coords When Hiding
in General geocaching topics
Posted
Fizzy, if you are going to quote someone then stop trying the smartarse trick and quote the whole context not your selectiveness.
One could come up with what ifs all day but one thing is for sure the track nature is not predictable but neither is it random.
BTW your also appear to have a little bit of selective tunnel vision especially when it comes to discussion regarding replies in the original context of the original poster.
You can continue to push crap up hill with a stick all day but at least make sure why your pushing it. Are you tring to say GPS position solutions are random, your a bigger fool than I thought.
Cheers, Kerry.