I aggree - if the satnav manufacturer states positional accuracy is only capable to the nearest 5 metres how can indicated speed be accurate? HGV drivers have to have their Tachos calibrated and sealed but even they are not classed as "accurate" for speed purposes only average speed. Aircraft do not rely on GPS for speed indication and use still use Inertial devices for precision navigation.
Commercial aricraft use the same standard re positioning navigational equipment, change has to be over time and the whole industry. Sat nav I understand will replace these systems in time and improve accuracy and safety. Military use sat nav more than civilains as it was a set-up for military use.
The whole reason that sat-nav was released to the world in general was due to an Aircraft safety issue. The new latest Giant Airbus relies on GPS for positiong due to its size.
NO one is saying that sat nav is 100% accurate, but it is far more accurate than what is generally available to the public and has been proven to be so by a variety of people in a business who rely on other equipment to assess speed.
Hi PC,
Got to concede that you are almost certainly right in your assertion that Satnav is very accurate in speed recording mode although I can't give you points for explaining why.
Done some more reading on the subject - well us retirees have to do something to keep our brains functioning! and it seems that although positional error does exist in the commonly available Satnavs it will not affect speed computation in "normal" operation.
The explanation is rather simple actually - basically if there is a positional error at a given point it will probably be caused atmospherically and if for example the error is 50 minutes N then it is likely to reproduce the error for a number of minutes rather than suddeny shift. So in simple terms if the sample time is every second (some satnavs are higher freq) the distance travelled will be unaffected by a positional error because as in our example you are travelling between two points that are both at 50 minutes north of their true position.
The digital electronics of the Satnav can then easily produce high speed accurate speed calculations from the distance travelled versas time base and as others have said display them to several decimal places if needed.
A small snag is the odd glitch when passing tall buidings, under bridges or map joins when a several samples are missed and the device catches up - some bikers have reported spikes of 300 MPH or more!
Ive just worked out that both my car speedos are 16% plus at 30MPH!
Got to concede that you are almost certainly right in your assertion that Satnav is very accurate in speed recording mode although I can't give you points for explaining why.
Done some more reading on the subject - well us retirees have to do something to keep our brains functioning! and it seems that although positional error does exist in the commonly available Satnavs it will not affect speed computation in "normal" operation.
The explanation is rather simple actually - basically if there is a positional error at a given point it will probably be caused atmospherically and if for example the error is 50 minutes N then it is likely to reproduce the error for a number of minutes rather than suddeny shift. So in simple terms if the sample time is every second (some satnavs are higher freq) the distance travelled will be unaffected by a positional error because as in our example you are travelling between two points that are both at 50 minutes north of their true position.
The digital electronics of the Satnav can then easily produce high speed accurate speed calculations from the distance travelled versas time base and as others have said display them to several decimal places if needed.
A small snag is the odd glitch when passing tall buidings, under bridges or map joins when a several samples are missed and the device catches up - some bikers have reported spikes of 300 MPH or more!
Ive just worked out that both my car speedos are 16% plus at 30MPH!
At least my ramblings inspired you to get somewhere where you realised I and others were right ;=)
I didn't want to confuse re atmospheric conditions and line of sight problems.
For most of us the only positional accuracy comes with the type of map in front of you on the unit.
My aerospace contact emailed his US sat nav contact, but the chap would not try and explain the workings of sat nav speed calculations re accuracy. But just said that the speed accuracy is not reliant on positional accuracy but that it as accurate as we in the public will generally get in our hands.
I didn't want to confuse re atmospheric conditions and line of sight problems.
For most of us the only positional accuracy comes with the type of map in front of you on the unit.
My aerospace contact emailed his US sat nav contact, but the chap would not try and explain the workings of sat nav speed calculations re accuracy. But just said that the speed accuracy is not reliant on positional accuracy but that it as accurate as we in the public will generally get in our hands.
Hello again Dave & PC,
I have to admit that I have been playing devils advocate on this one, but my stance has not been entirely unreasonable. GPS and Sat Nav (I deliberately make the distinction between the two) rely on a mammoth amount of technology, and it only needs one part of the chain to become faulty or have 'bugs' to potentially make the final display unreliable.
I am still sceptical about the claims of 'accurate' positional resolutions, and if there is still the potential for positional error, for what ever reason, then the calculation of speed must be flawed because of the error in the raw data. I do not accept that positional errors would simply be a constant offset, I believe there is strong potential for a random offset, which could at worst add or subtract double the offset to the distance between two successive sample points.
Even given the military background to GPS, in the theatre of war an error of 5 meters for an air bomb or missile would still considered precision. There are surveying systems that utilise GPS, for accuracies in the region of mm, but the acquisition time to achieve a high quality fix is in the order of minutes - which is not a practical solution for moving vehicles or missiles!
A 5 meter positional error for an aircraft in flight, or the difference between a speed reading of 550 and 551 Knots would not pose a significant hazard or problem, but on the ground where the Government is keen to see speed limits obeyed then a 1mph error is all it takes.
I have searched and not found any information regarding the error related to the uncertainty of the least significant digit. If any of you have found sites please let me know their web addresses.
From practical use of a GPS and Sat Nav, I do find that the speeds shown on the whole are very believable. But equally I have tried to use a hand held Garmin Etrek GPS for plotting out a large camping site, and found that LSD 2 digits of GB OS Grid Lat & Long readings were continually changing. This made the plot inaccurate.
Even PC has agreed that on occasions some errors can occur.
In conclusion, I agree that GPS tends to give relatively good speed accuracy, though you must be aware that some inaccuracy can occur and for that reason GPS and Sat Nav are not infallible
By contrast the error on a car speedo will remain fairly constant and as such you can compensate with quite reliable and predictable results.
Unless I get any more traceable evidence I am unlikely to add more to this debate.
I have to admit that I have been playing devils advocate on this one, but my stance has not been entirely unreasonable. GPS and Sat Nav (I deliberately make the distinction between the two) rely on a mammoth amount of technology, and it only needs one part of the chain to become faulty or have 'bugs' to potentially make the final display unreliable.
I am still sceptical about the claims of 'accurate' positional resolutions, and if there is still the potential for positional error, for what ever reason, then the calculation of speed must be flawed because of the error in the raw data. I do not accept that positional errors would simply be a constant offset, I believe there is strong potential for a random offset, which could at worst add or subtract double the offset to the distance between two successive sample points.
Even given the military background to GPS, in the theatre of war an error of 5 meters for an air bomb or missile would still considered precision. There are surveying systems that utilise GPS, for accuracies in the region of mm, but the acquisition time to achieve a high quality fix is in the order of minutes - which is not a practical solution for moving vehicles or missiles!
A 5 meter positional error for an aircraft in flight, or the difference between a speed reading of 550 and 551 Knots would not pose a significant hazard or problem, but on the ground where the Government is keen to see speed limits obeyed then a 1mph error is all it takes.
I have searched and not found any information regarding the error related to the uncertainty of the least significant digit. If any of you have found sites please let me know their web addresses.
From practical use of a GPS and Sat Nav, I do find that the speeds shown on the whole are very believable. But equally I have tried to use a hand held Garmin Etrek GPS for plotting out a large camping site, and found that LSD 2 digits of GB OS Grid Lat & Long readings were continually changing. This made the plot inaccurate.
Even PC has agreed that on occasions some errors can occur.
In conclusion, I agree that GPS tends to give relatively good speed accuracy, though you must be aware that some inaccuracy can occur and for that reason GPS and Sat Nav are not infallible
By contrast the error on a car speedo will remain fairly constant and as such you can compensate with quite reliable and predictable results.
Unless I get any more traceable evidence I am unlikely to add more to this debate.
sat nav, versus car speedo?
john.
the sat nav wins. you see firstly a manufacturer must be sure
his speedo does not under read, so they will always be optimestic.
what about tyres? rolling radius? these will for sure effect your cars speedo. eg, a 205x 65x16 tyre made by different manufactures can have a difference of 1.5% and your speedo cannot
tell the diffeence, nor can it tell if the tyre is hot or cold
or new or warn.
these are only very small differences, but they do add up.
the sat nav on the otherhand does not use contact with the road as such, so these differences do not effect it, nor do they have
to build in a safety factor for being accurate, or rather making
sure they do not under read.
so in the majority of situations the sat nav wins for accuracy.
As I drive I have have sat nav and speedo in front of me, mine and other sat navs show my speedo as being 3mph out and my wifes cars as 3mph out, this is the same across consistent speeds from 10 to 140 mph+ not on public road.
You are actually wrong I believe that I'm told even speed cameras and detection limit still gives a small leeway for error.
All sat navs can't be wrong that my speedo is 3mph out and others much the same what ever science or error factors might creep in.
It has been suggested to e that the way the Sat Nav Calculates speed still it still has little basis on positioning. I may be lost but I still know that I'm doing 70, much the same with the sat nav, it night get positioning slightly out but I and many others do not get a sudden difference greater difference between the cars speedo and the sat nav.
When in navigation mode my Garmin units do not show the speed, but I know from trial and error that my speedo is 3 mph out.
No matter what the science involved is and what the hard weat or soft wear errors might be, all reported evidence points to better than 1% accuracy re speed and you can bet your bottom dollar some yank would have sued if that were not the case.
My friend who drives an old Toyota uased to accuse me of speeding when I got sat nav we put it in his car and knowing that nine was 3mph out with my speedo on 73 and the sat nav in his car on 70 and his speed reading 65 we kept station.
My wife has a small Garmin in her car and we get the same.
When you talk about positional error, you are also basing it on information traveling between a number of satellites to your sat nav. As you need at least 6 or 7 satellite locks to get 3 - 5 metre accuracy that involves signals traveling between 72000 miles or so wand the codes being crunched to give you position in less than a second. Now even if you were 100 feet out I don't think the error factors decimal points would fit on my fairly basic calculator.
It has been said that the sat nav speed is most accurate on long straight level roads, I don't think I or anyone else is suggesting that you blank out your speedo and watch the sat nav reading, just that you know that the error on sat nav makes it generally more accurate than the average speedo that is made inaccurate in the first place.
A good question and some great answers, so I'll throw a bit more in.
Vehicle manufacturers have to ensure that the speedo never reads fast - so in addition to the accuracy of the speedo they have to take the worst case scenario with the tyres.
It's all about what's called 'rolling radius' as a tyre that is almost at the end of it's life (say 1.6mm tread left) will rotate faster than a new tyre (say an additional 5mm).
On the flat +/- sat navs are spot on - next time someones going to have a new set of tyres (or front pair if it's front wheel drive) do a speed check against the sat nav on the way to the tyre garage and then another on the way back - new tyres will read a slower speed.
Vehicle manufacturers have to ensure that the speedo never reads fast - so in addition to the accuracy of the speedo they have to take the worst case scenario with the tyres.
It's all about what's called 'rolling radius' as a tyre that is almost at the end of it's life (say 1.6mm tread left) will rotate faster than a new tyre (say an additional 5mm).
On the flat +/- sat navs are spot on - next time someones going to have a new set of tyres (or front pair if it's front wheel drive) do a speed check against the sat nav on the way to the tyre garage and then another on the way back - new tyres will read a slower speed.
Dear PC,
I do not disagree that you find a consistent difference between your car speedometer and sat navs. However it is scientifically dangerous to assume that just because the same result is achieved many times with the same method that it is in fact the truth, especially when there are so many potential causes of errors in the system. For the result to have real authority it needs to be corroborated against traceable standards by an alternative methods. Perhaps the question should be 'Can all sat Navs be right'?
Perhaps I did not explain my reference to speed cameras very well. The camera will have some defined detection limit of X, If you are travelling at X+1 you will trigger the camera. I happened to use 30mph as an example as some forces are advocating zero tolerance speed limits.
Now you set your speed according to the Sat Nav. Then because of a downhill section or you apply slightly more pressure to the accelerator, or what ever your speed begins to creep upwards:
30.1
30.2
30.3
30.4
30.5
30.6
30.7
30.8
30.9
31.0
At what real speed does the sat nav change from 30 to 31? you could be travelling at 30.9 and still be showing only 30 This is part of the error that all analogue to digital conversions have and leads to uncertainty in the least significant digit, on top of any other inaccuracies.
If the camera is set to 30mph, then your 30.9 could be enough to trigger it. Exactly the same scenario can occur at any given speed.
Sorry to be boring but speed is defined as distance divided by time. I do not know of any other fundamental method of establishing speed. GPS only provides positional and time information. It is down to other software to calculate distance, and therefore speed. The accuracy of the speed indication will be a function of the errors in establishing the position of the two samples and the accuracy of the time base used to determine the interval time.
GPS takes time to establish a positional fix. The more time available the more accurate the fix, conversely the less time the more inaccurate the fix will be. As an indication of the time, surveyors need several minutes for their sophisticated GPS equipment to establish a high probability of a cm accuracy fix.
That is why automotive GPS which cycles typically at 1fix/sec. Is unlikely to achieve better than a +/- 1m fix and could be as much as +/-5 m. These errors are not necessarily all biased in the same direction or size, and so it is possible for the first fix of a pair to be under read by 5m and the second fix to be over by 5m, giving a potential total error for two consecutive samples of 10M.
However this margin of error only applies to the first and last fixes in a string, because all intermediate positional errors will virtually cancel out. So for a journey of 1 second the potential error for a car is alarming, but as the total distance increases the combined errors of the first and last readings become a far smaller percentage of the whole journey and will lead to a vastly improved calculation of distance travelled and thus speed. The problem is that it only updates the display infrequently
There are other techniques that such a using a rolling average which can allow an improved reading of speed which whilst delayed by a few seconds as the number of samples is built up will track speed changes at a rate that is of use to drivers. I believe that this is process used by most Sat Navs to display speed to reduce some of the error.
The error factor we are dealing with here is the ability for a GPS unit to interrogate the satellites transmission s and its ability to fix its position at ant time. The error has been quoted as 1 to 5m. And whilst that is a very small percentage of the signals total journey, it is a significant error on the ground when calculating fix to fix speeds.
I have never said that you should blank out your speedo, and I have never suggested that you have suggested it. Though with you certainty in the accuracy of Sat Nav speeds, you could argue a case, or perhaps you do have some nagging doubts about them?
I do agree that on the whole Sat Nav is likely to give a closer to the truth speed reading than the majority of standard car speedo's, BUT I do have doubts in some circumstances and will therefore continue to use them with some caution.
I do not disagree that you find a consistent difference between your car speedometer and sat navs. However it is scientifically dangerous to assume that just because the same result is achieved many times with the same method that it is in fact the truth, especially when there are so many potential causes of errors in the system. For the result to have real authority it needs to be corroborated against traceable standards by an alternative methods. Perhaps the question should be 'Can all sat Navs be right'?
Perhaps I did not explain my reference to speed cameras very well. The camera will have some defined detection limit of X, If you are travelling at X+1 you will trigger the camera. I happened to use 30mph as an example as some forces are advocating zero tolerance speed limits.
Now you set your speed according to the Sat Nav. Then because of a downhill section or you apply slightly more pressure to the accelerator, or what ever your speed begins to creep upwards:
30.1
30.2
30.3
30.4
30.5
30.6
30.7
30.8
30.9
31.0
At what real speed does the sat nav change from 30 to 31? you could be travelling at 30.9 and still be showing only 30 This is part of the error that all analogue to digital conversions have and leads to uncertainty in the least significant digit, on top of any other inaccuracies.
If the camera is set to 30mph, then your 30.9 could be enough to trigger it. Exactly the same scenario can occur at any given speed.
Sorry to be boring but speed is defined as distance divided by time. I do not know of any other fundamental method of establishing speed. GPS only provides positional and time information. It is down to other software to calculate distance, and therefore speed. The accuracy of the speed indication will be a function of the errors in establishing the position of the two samples and the accuracy of the time base used to determine the interval time.
GPS takes time to establish a positional fix. The more time available the more accurate the fix, conversely the less time the more inaccurate the fix will be. As an indication of the time, surveyors need several minutes for their sophisticated GPS equipment to establish a high probability of a cm accuracy fix.
That is why automotive GPS which cycles typically at 1fix/sec. Is unlikely to achieve better than a +/- 1m fix and could be as much as +/-5 m. These errors are not necessarily all biased in the same direction or size, and so it is possible for the first fix of a pair to be under read by 5m and the second fix to be over by 5m, giving a potential total error for two consecutive samples of 10M.
However this margin of error only applies to the first and last fixes in a string, because all intermediate positional errors will virtually cancel out. So for a journey of 1 second the potential error for a car is alarming, but as the total distance increases the combined errors of the first and last readings become a far smaller percentage of the whole journey and will lead to a vastly improved calculation of distance travelled and thus speed. The problem is that it only updates the display infrequently
There are other techniques that such a using a rolling average which can allow an improved reading of speed which whilst delayed by a few seconds as the number of samples is built up will track speed changes at a rate that is of use to drivers. I believe that this is process used by most Sat Navs to display speed to reduce some of the error.
The error factor we are dealing with here is the ability for a GPS unit to interrogate the satellites transmission s and its ability to fix its position at ant time. The error has been quoted as 1 to 5m. And whilst that is a very small percentage of the signals total journey, it is a significant error on the ground when calculating fix to fix speeds.
I have never said that you should blank out your speedo, and I have never suggested that you have suggested it. Though with you certainty in the accuracy of Sat Nav speeds, you could argue a case, or perhaps you do have some nagging doubts about them?
I do agree that on the whole Sat Nav is likely to give a closer to the truth speed reading than the majority of standard car speedo's, BUT I do have doubts in some circumstances and will therefore continue to use them with some caution.
John, I'm no scientist. But sat nav signals are received at the speed of light. And I'm not suggesting that you replace looking at speedo with your Sat Nav.
Betwen family and worl associates we have access to numerous Sat Navs. The one common factor they share is speed readings. A close associates works in a business where
Betwen family and worl associates we have access to numerous Sat Navs. The one common factor they share is speed readings. A close associates works in a business where
Dear PC,
I have tried to be a clear in my description of the potential error and how it can occur, but you seem to miss-interpret my examples as real life situations.
Yes the satellite transmissions will travel at the speed of light, and the system uses that constant, but the speed of the signal is not in its self important, but the time delay between signals from different satellites reaching the GPS receiver is. By knowing the relative position of the satellites the time difference between the signals will allow the GPS to calculate the relative distances from each of the satellites and by triangulation can calculate a fix.
Fact - if a GPS generates positional errors when making a fix, The any speed calculation based on that fix will its self be erroneous.
It is also a fact that some police forces have suggested a zero tolerance should be adopted to speed limits. As has been stated, the EU construction regulations only permit car speedometers to show the accurate speed or greater than. This would enable cameras to be set at precise limits, because no car should indicate less than its real speed.
Until zero tolerance is adopted than the margin or leeway on camera trigger speeds is not prescribed in any law or regulations but is down to the local authorities, and the margin varies across the country.
Each camera is set to trigger at some speed which we will call X. whilst it may not be equal to the speed limit it is still a precise setting which, if a vehicle exceeds that speed the camera is triggered.
The point is that if you have knowledge of what a particular camera's trigger speed (x) is, you may be tempted to use your 'accurate' Sat Nav to match that speed. And again I make the point that if the sat nav is indicating any speed to 2 significant figures (i.e. 0 to 99 mph) what is your actual speed? For any indicated speed the digitisation error, and the truncating of the speed could put you speed anywhere between +/- 1 of the indication. Without the sat Nav Manufacture defining the the way the decimal parts of the speed are truncated you have to assume the worst case.
So if you know a camera is set to for example 35mph, and your sat nav indicates you are travelling at 35mph, with the digitisation error you could be travelling as slow as 34.1 and 35.9 , which is sufficient to trigger the camera. The same principal applies to any speed you care to choose.
These digitisation errors are on top of the positional fix errors that occur - so it is foolhardy to accept as beyond doubt the sat nav display of speed.
Now if you friends find sat nav to be as accurate as their own conventional measurements, ask them why they don't simply use SAT Nav and dispose of all their expensive traceable instrument's . - probably, it is not accepted as authoritative evidence because the instrumentation cannot be verified through national standards. As it is dependant on the motion of the satellites and as a result the number and identity of satellites that are in view continually change, there are too many variables to be able with absolute confidence to rely on the indications of a SAT Nav as a definitive indicator.
Please continue to use Sat Nav as a way to more accurately establish your speed, But understand that there are potential errors in the reading it gives. do allow for the digitisation error not assume it is correct all the time, and also.
This is my last word on this thread.
I have tried to be a clear in my description of the potential error and how it can occur, but you seem to miss-interpret my examples as real life situations.
Yes the satellite transmissions will travel at the speed of light, and the system uses that constant, but the speed of the signal is not in its self important, but the time delay between signals from different satellites reaching the GPS receiver is. By knowing the relative position of the satellites the time difference between the signals will allow the GPS to calculate the relative distances from each of the satellites and by triangulation can calculate a fix.
Fact - if a GPS generates positional errors when making a fix, The any speed calculation based on that fix will its self be erroneous.
It is also a fact that some police forces have suggested a zero tolerance should be adopted to speed limits. As has been stated, the EU construction regulations only permit car speedometers to show the accurate speed or greater than. This would enable cameras to be set at precise limits, because no car should indicate less than its real speed.
Until zero tolerance is adopted than the margin or leeway on camera trigger speeds is not prescribed in any law or regulations but is down to the local authorities, and the margin varies across the country.
Each camera is set to trigger at some speed which we will call X. whilst it may not be equal to the speed limit it is still a precise setting which, if a vehicle exceeds that speed the camera is triggered.
The point is that if you have knowledge of what a particular camera's trigger speed (x) is, you may be tempted to use your 'accurate' Sat Nav to match that speed. And again I make the point that if the sat nav is indicating any speed to 2 significant figures (i.e. 0 to 99 mph) what is your actual speed? For any indicated speed the digitisation error, and the truncating of the speed could put you speed anywhere between +/- 1 of the indication. Without the sat Nav Manufacture defining the the way the decimal parts of the speed are truncated you have to assume the worst case.
So if you know a camera is set to for example 35mph, and your sat nav indicates you are travelling at 35mph, with the digitisation error you could be travelling as slow as 34.1 and 35.9 , which is sufficient to trigger the camera. The same principal applies to any speed you care to choose.
These digitisation errors are on top of the positional fix errors that occur - so it is foolhardy to accept as beyond doubt the sat nav display of speed.
Now if you friends find sat nav to be as accurate as their own conventional measurements, ask them why they don't simply use SAT Nav and dispose of all their expensive traceable instrument's . - probably, it is not accepted as authoritative evidence because the instrumentation cannot be verified through national standards. As it is dependant on the motion of the satellites and as a result the number and identity of satellites that are in view continually change, there are too many variables to be able with absolute confidence to rely on the indications of a SAT Nav as a definitive indicator.
Please continue to use Sat Nav as a way to more accurately establish your speed, But understand that there are potential errors in the reading it gives. do allow for the digitisation error not assume it is correct all the time, and also.
This is my last word on this thread.
Facebook
Twitter