SLOW DOWN

[GUMPY]

3mph, that's stationary 😱

Normal cruising speed is 5+mph, I might back it off to below 4mph when passing boats but then I may not🤭

[MtB]

8 minutes ago, Loddon said:

3mph, that's stationary 😱

Normal cruising speed is 5+mph, I might back it off to below 4mph when passing boats but then I may not🤭

 

 

Is that what your GPS is telling you??!!

 

[GUMPY]

8 minutes ago, MtB said:

 

 

Is that what your GPS is telling you??!!

 

Which GPS?

The Garmin tells me that I travel at 5mph the 3mph comment was humour. I never check the other ones.

Indicated 5mph is 1400 rpm on a nice wide deep River.

 

Edited May 7, 2022 by Loddon

[IanD]

27 minutes ago, David Mack said:

Indeed, but ticket matters apart, knowing your speed in a car is 70 mph +/- 3mph is fine. Being told your boat is doing 3mph +/- 3 mph is useless!

GPS measures position, speed is indeed the derivative of position, but accuracy of speed (e.g. 0.1mph) is independent of speed -- if the GPS reading is accurate to within 0.1mph, it's be that accurate at 0mph or 3mph or 70mph.

 

Unless of course it switches/adapts algorithms as I described might happen, in which case there has to be less smoothing at high speeds to track fast changes of direction, which means more reading noise at high speeds and less at low speeds... 😉

[nicknorman]

On 07/05/2022 at 12:28, IanD said:

GPS measures position, speed is indeed the derivative of position, but accuracy of speed (e.g. 0.1mph) is independent of speed -- if the GPS reading is accurate to within 0.1mph, it's be that accurate at 0mph or 3mph or 70mph.

 

Unless of course it switches/adapts algorithms as I described might happen, in which case there has to be less smoothing at high speeds to track fast changes of direction, which means more reading noise at high speeds and less at low speeds... 😉

As I understand it there are a couple of ways of measuring velocity with gps, one obviously based on rate of change of position, but perhaps a better way is directly, using the Doppler shift of the satellite signals, which of course gives an instantaneous velocity rather than one relying on comparing adjacent fixes.

[MtB]

1 hour ago, nicknorman said:

As I understand it there are a couple of ways of measuring velocity with gps, one obviously based on rate of change of position, but perhaps a better way is directly, using the Doppler shift of the satellite signals, which of course gives an instantaneous velocity rather than one relying on comparing adjacent fixes.

 

 

I'd imagine the Doppler shift at 3mph would be microscopically small so just as prone to inaccuracy as rate of change of position.

[IanD]

2 hours ago, nicknorman said:

As I understand it there are a couple of ways of measuring velocity with gps, one obviously based on rate of change of position, but perhaps a better way is directly, using the Doppler shift of the satellite signals, which of course gives an instantaneous velocity rather than one relying on comparing adjacent fixes.

AFAIK all GPS recievers use position as the primary variable, everything else is worked out from this.

 

Edited May 8, 2022 by IanD

[nicknorman]

2 hours ago, MtB said:

 

 

I'd imagine the Doppler shift at 3mph would be microscopically small so just as prone to inaccuracy as rate of change of position.

Not really. All (most) gps satellites broadcast on the same frequency. The carrier is extremely high frequency around 1.5GHz and the modulated data is very low frequency (50bits/sec). The overall bandwidth of the signal is thus very narrow. Very narrow filters are used to select the wanted satellite data from the others. But the Doppler shift means that the centre frequency varies by many, many times the bandwidth of the signal, so the filter frequency has to track the satellite’s Doppler shift, adjusting all the time to keep the satellite in tune.

 

After all, a policeman with his radar gun can tell your speed to within an Mph or so using Doppler.

 

This is why, especially with older GPSs, they could take several minutes to get a fix when turned on from “cold” or if they had been moved a long way when switched off. They would give a message like “searching the sky”. This made it sound a bit like they had some sort of directional aerial, but of course they didn’t. Searching the sky involved slowly sweeping the filter frequency until a valid satellite signal was found, tracking it (frequency-wise) then moving on to find the next satellite. The data about satellites’ position and velocity in the future is downloaded and kept in non-volatile memory in a modern GPS, which also has its own clock, so when you switch it on it generally has a very good idea which satellites are in view and what frequency offset (Doppler shift) they will have, so the receiver can get a fix in a few seconds.

1 hour ago, IanD said:

AFAIK all GPS recievers use position as the primary variable, everything else is worked out from this.

 

I disagree. This quote from the article below:

“GPS and other global navigation satellite systems use the Doppler shift of the received carrier frequencies to determine the velocity of a moving receiver. Doppler-derived velocity is far more accurate than that obtained by simply differencing two position estimates.”

 

https://www.gpsworld.com/gnss-systemalgorithms-methodsinnovation-doppler-aided-positioning-11601/

 

Edited May 8, 2022 by nicknorman

[TheBiscuits]

1 hour ago, IanD said:

AFAIK all GPS recievers use position as the primary variable, everything else is worked out from this.

 

Ah   I always understood them to use a timestamp and nothing else.

 

I'll not argue with someone who has designed GPS receiver chipsets, but I thought the satellites were nothing but fancy clock repeaters.

 

Is that not the case?

[IanD]

11 hours ago, nicknorman said:

Not really. All (most) gps satellites broadcast on the same frequency. The carrier is extremely high frequency around 1.5GHz and the modulated data is very low frequency (50bits/sec). The overall bandwidth of the signal is thus very narrow. Very narrow filters are used to select the wanted satellite data from the others. But the Doppler shift means that the centre frequency varies by many, many times the bandwidth of the signal, so the filter frequency has to track the satellite’s Doppler shift, adjusting all the time to keep the satellite in tune.

 

After all, a policeman with his radar gun can tell your speed to within an Mph or so using Doppler.

 

This is why, especially with older GPSs, they could take several minutes to get a fix when turned on from “cold” or if they had been moved a long way when switched off. They would give a message like “searching the sky”. This made it sound a bit like they had some sort of directional aerial, but of course they didn’t. Searching the sky involved slowly sweeping the filter frequency until a valid satellite signal was found, tracking it (frequency-wise) then moving on to find the next satellite. The data about satellites’ position and velocity in the future is downloaded and kept in non-volatile memory in a modern GPS, which also has its own clock, so when you switch it on it generally has a very good idea which satellites are in view and what frequency offset (Doppler shift) they will have, so the receiver can get a fix in a few seconds.

I disagree. This quote from the article below:

“GPS and other global navigation satellite systems use the Doppler shift of the received carrier frequencies to determine the velocity of a moving receiver. Doppler-derived velocity is far more accurate than that obtained by simply differencing two position estimates.”

 

https://www.gpsworld.com/gnss-systemalgorithms-methodsinnovation-doppler-aided-positioning-11601/

 

 

That's not how GPS works either, the signal is not narrowband, it's spread-spectrum and fills the entire GPS band, and is extracted from far below the noise floor by correlation -- the raw bit rate is indeed low (50Hz) but the modulated bit rate after 1023b sequence multiplication (different for each satellite) is much higher (1.023MHz). The problem for the receiver is that is has to search for and lock onto all these spread-spectrum signals buried in the noise, and since the satellites are moving quite rapidly the correlation sequences are moving in time (and frequency, due to Doppler shift) so are tricky to find -- it's like looking for a needle in a haystack, once you've found it (locked on) things are fine, the challenge is locking on in the first place.

 

The description of locking on is correct, the almanac and ephemeris data is stored in the receiver. Ones with internet access can also download predicted satellite data for the near future to enable faster lock, as well as getting rough position data from other sources like WiFi or cellular signals. Or since the position of the satellites is predictable (orbits don't move much), it's possible to calculate where they'll be at a given time so long as you know what the time is.

 

I don't know if the Doppler shift method is actually used by modern chip sets or not, the paper just shows that it *could* be used -- don't forget that the main purpose of most GPS receivers is to provide position data, speed is a secondary requirement in most cases, so if this technique is used at all in real life it might only pop up in specialised receivers intended for speed measurement, not those in mobile handsets where the manufacturer can't be bothered with the extra algorithm hassle for no benefit in selling more phones -- or maybe they do, there's no way of finding out. Nowadays all the GPS function is integrated into the phone SoC (e.g. from Qualcomm or Mediatek or Apple or Samsung or...), it's not a separate chipset like it used to be where such a feature might sell more chipsets -- which could also be what is done in specialised GPS receivers, separate GPS chips do still exist and might well use Doppler shift.

 

There's a lot of useful information in the answers here...

 

https://www.quora.com/If-GPS-works-by-sensing-Doppler-shifts-of-satellite-signals-why-doesnt-it-get-scrambled-when-we-move-quickly-as-in-aircraft

 

"Doppler shift between the user’s receiver and spacecraft can also be used. For this user’s device must be equipped with frequency shift detection necessary signal processing methods."

 

Position estimation relies on knowing exactly where the satellites are at a given instant in time, each one broadcasts this time as part of the data, and the GPS receiver uses time differences between the received signals and 3D trigonometry to work out where it is -- which is why signals from at least four satellites are needed. All the GPS satellites have synchronised atomic clocks (within a few nanoseconds) in them, all locked back to the global NIST reference clock IIRC.

Edited May 9, 2022 by IanD

[Mike Todd]

16 hours ago, nicknorman said:

As I understand it there are a couple of ways of measuring velocity with gps, one obviously based on rate of change of position, but perhaps a better way is directly, using the Doppler shift of the satellite signals, which of course gives an instantaneous velocity rather than one relying on comparing adjacent fixes.

I refer to an earlier post: in developing an application based on a USB GPS unit (I actually tried two very different ones, both USB and Bluetooth) and studied the actual data stream from the unit. Speed is especially unstable at boat speeds. AFAIK, they did not stream any Doppler data.

[nicknorman]

20 hours ago, IanD said:

 

I don't know if the Doppler shift method is actually used by modern chip sets or not, the paper just shows that it *could* be used

 

No, the paper says it *is* used for velocity. The "could" bit is the purpose of the paper in that when you know the velocity accurately, that info could be used to refine the position. By some sort of integration presumably. More or less the reverse of working out the velocity by differentiating position. Whether it *is* used for velocity in every case, or just some cases, I don't know. Anyway my original point was that there are 2 ways a GPS receiver can work out velocity, and using doppler is the more accurate. Having observed my phone measuring boat speed, it looks like it uses differential position rather than doppler. But a better receiver would use doppler.

 

Yes agreed on the actual bandwidth, but the receiver still needs to track the doppler shift to remain in tune. This is why aviation GPSs specify a maximum velocity (maximum doppler frequency offset) and a maximum acceleration in 3D (Maximum rate of change of doppler frequency offset to retain lock)

[IanD]

1 hour ago, nicknorman said:

No, the paper says it *is* used for velocity. The "could" bit is the purpose of the paper in that when you know the velocity accurately, that info could be used to refine the position. By some sort of integration presumably. More or less the reverse of working out the velocity by differentiating position. Whether it *is* used for velocity in every case, or just some cases, I don't know. Anyway my original point was that there are 2 ways a GPS receiver can work out velocity, and using doppler is the more accurate. Having observed my phone measuring boat speed, it looks like it uses differential position rather than doppler. But a better receiver would use doppler.

 

Yes agreed on the actual bandwidth, but the receiver still needs to track the doppler shift to remain in tune. This is why aviation GPSs specify a maximum velocity (maximum doppler frequency offset) and a maximum acceleration in 3D (Maximum rate of change of doppler frequency offset to retain lock)

 

The first paragraph is arguing about how many angels can dance on the head of a pin; there's no way to know if doppler shift is actually used in a given GPS receiver unless the manufacturer specifically says that it is. I can tell you that it wasn't used in the chipset I designed, and I suspect not in any of the competitors we benchmarked against (this was in the days of discrete GPS chipsets before they got integrated into phone SoC). You say your phone seems to use position (which is what I would expect) and then that "a better receiver would use Doppler" -- which is just speculation, it could and might do but I haven't found any claims to do this, certainly not in standard GPS receivers.

 

It's like quoting a paper from McLaren showing that gold foil is the best solution for heat reflection in an engine compartment (it is, they used it in the F1), and then saying that this means that "a better car would use this" -- yes it's possible in theory, but it doesn't happen in practice because of demand and cost.

 

There is no theoretical upper limit to the amount of Doppler shift that can be tracked, because due to the 3D trigonometry used this is different anyway for every visible satellite and direction of travel (in three dimensions). Civilian GPS receivers specify a maximum velocity (and acceleration?) because this is a legal COCOM requirement to stop them being used in military applications like smart munitions and missiles -- if velocity exceeds 1000kts (1150mph/1850kph) or height exceeds 18000m (59000ft) the receiver has to return null position location. This limit is hardwired into the design, or in some cases the firmware -- which then has to be locked down to prevent unauthorised modifications by anyone other than the device manufacturer.

 

Go on, ask me how I know this... 😉

[Alan de Enfield]

20 minutes ago, IanD said:

Go on, ask me how I know this... 😉

 

Google knows it all.

[IanD]

2 minutes ago, Alan de Enfield said:

 

Google knows it all.

Google doesn't know the ins and outs of how the COCOM rules on hardware/firmware are enforced though, because Google didn't have to sign the documents guaranteeing to follow them... 😉

Edited May 10, 2022 by IanD

[nicknorman]

39 minutes ago, IanD said:

 

The first paragraph is arguing about how many angels can dance on the head of a pin; there's no way to know if doppler shift is actually used in a given GPS receiver unless the manufacturer specifically says that it is. I can tell you that it wasn't used in the chipset I designed, and I suspect not in any of the competitors we benchmarked against (this was in the days of discrete GPS chipsets before they got integrated into phone SoC). You say your phone seems to use position (which is what I would expect) and then that "a better receiver would use Doppler" -- which is just speculation, it could and might do but I haven't found any claims to do this, certainly not in standard GPS receivers.

 

It's like quoting a paper from McLaren showing that gold foil is the best solution for heat reflection in an engine compartment (it is, they used it in the F1), and then saying that this means that "a better car would use this" -- yes it's possible in theory, but it doesn't happen in practice because of demand and cost.

 

There is no theoretical upper limit to the amount of Doppler shift that can be tracked, because due to the 3D trigonometry used this is different anyway for every visible satellite and direction of travel (in three dimensions). Civilian GPS receivers specify a maximum velocity (and acceleration?) because this is a legal COCOM requirement to stop them being used in military applications like smart munitions and missiles -- if velocity exceeds 1000kts (1150mph/1850kph) or height exceeds 18000m (59000ft) the receiver has to return null position location. This limit is hardwired into the design, or in some cases the firmware -- which then has to be locked down to prevent unauthorised modifications by anyone other than the device manufacturer.

 

Go on, ask me how I know this... 😉

 

I know it is very important to you to always be right and to hold the power in any discussion, but I will just mention that I simply said that there were 2 ways to measure velocity using gps, one being doppler, and I provided evidence to support that point. I did not say, nor do I think, that your average phone gps uses doppler for velocity. So your campaign to discredit my point seems a bit OTT although entirely typical.

Yes I am sure that there are ultimate legal limits to velocity and height for the reason you mention, but that doesn't mean that a gps receiver has to be able to reach those limits. I haven't looked recently but in the earlier days of GA (leisure aviation) GPSs, a max velocity was quoted, often 250kts, and a maximum acceleration quoted, often around 5g. These limits arose simply because it is easier to make a gps engine that doesn't require too much frequency offset and whose tracking algorithms can be filtered to retain tracking more easily at the expense of some agility.

[IanD]

24 minutes ago, nicknorman said:

 

I know it is very important to you to always be right and to hold the power in any discussion, but I will just mention that I simply said that there were 2 ways to measure velocity using gps, one being doppler, and I provided evidence to support that point. I did not say, nor do I think, that your average phone gps uses doppler for velocity. So your campaign to discredit my point seems a bit OTT although entirely typical.

Yes I am sure that there are ultimate legal limits to velocity and height for the reason you mention, but that doesn't mean that a gps receiver has to be able to reach those limits. I haven't looked recently but in the earlier days of GA (leisure aviation) GPSs, a max velocity was quoted, often 250kts, and a maximum acceleration quoted, often around 5g. These limits arose simply because it is easier to make a gps engine that doesn't require too much frequency offset and whose tracking algorithms can be filtered to retain tracking more easily at the expense of some agility.

 You're first line is a bit rich, given that you're the one pontificating about something you obviously know little about... 😉

 

Second point -- I suggest you look recently, then, because things have changed a lot since then. I was working on GPS in the early 2000s and even then there were no such velocity limits, this is trivial to deal with. Acceleration can mess up the algorithms which attempt to predict where the vehicle is going next, but usually this then just switches it into a less accurate mode which only uses the position data instead of motion estimation -- or the algorithm coefficients adapt to this instead of simply switching modes.

 

Third point -- having done it, it's not "easier to make a GPS engine that doesn't require too much frequency offset", the effort is exactly the same. As stated earlier, tracking accuracy vs. speed to react to changes is an algorithm choice, but is not affected by velocity/frequency offset. High speed can increase acquisition time from cold because the search space is bigger, but many methods are available and used to speed this up.

 

This used to be a problem in the very early GPS days with chipsets with limited processing power, but isn't nowadays because the signal searching is many times faster because they run lots of acquisition processes in parallel -- they used to have one bloke looking for a needle in a haystack, today they have hundreds... 🙂

 

Edited May 10, 2022 by IanD

[nicknorman]

4 minutes ago, IanD said:

Second point -- I suggest you look recently, then. I was working on GPS in the early 2000s and there were no such velocity limits, this is trivial to deal with. Acceleration can mess up the algorithms which attempt to predict where the vehicle is going next, but usually this then just switches it into a less accurate mode which only uses the position data instead of motion estimation -- or the algorithm coefficients adapt to this instead of simply switching modes.

 

Third point -- having done it, it's not "easier to make a GPS engine that doesn't require too much frequency offset", the effort is exactly the same. As stated earlier, tracking accuracy vs. speed to react to changes is an algorithm choice, but is not affected by velocity/frequency offset. High speed can increase acquisition time from cold because the search space is bigger, but many methods are available and used to speed this up.

 

So if it is not easier to design a GPS engine with velocity and acceleration limits below the legal limit, why did many gps receiver designers do so? I am sure that with modern technological advances the difference becomes trivial, but my point (aimed at MtB if you recall - it's not all about you) goes back to the fact that the receiver has to track the doppler shift of the satellite signal and thus the doppler shift is significant.

[IanD]

23 minutes ago, nicknorman said:

So if it is not easier to design a GPS engine with velocity and acceleration limits below the legal limit, why did many gps receiver designers do so? I am sure that with modern technological advances the difference becomes trivial, but my point (aimed at MtB if you recall - it's not all about you) goes back to the fact that the receiver has to track the doppler shift of the satellite signal and thus the doppler shift is significant.

 

I explained that in my last post, but you obviously didn't catch it.

 

When trying to acquire satellites the receiver has to search through all the possible cases for delay and Doppler shift to lock on to each one. The bigger the speed range the bigger the Doppler shift, which means the search space is larger. If it's too large and the search takes too long then it can take a very long time or even fail to ever lock on.

 

This was what used to happen in the "old days" when processing power was limited by the processing speed in the chipset, and there was only one receiver searching for signals -- or at most a few, because more than this cost too much and consumed too much power -- so one "fix" (a bodge, really...) was to limit the Doppler search range by restricting maximum velocity.

 

Nowadays with modern chip technology the cost and power of each virtual receiver channel -- remember these are all digital signal processing, there's only one antenna input -- is absolutely tiny, transistor count has gone up by maybe 1000x since the design I was talking about, see attached plot. So they simply run a massive number (tens? hundreds? thousands?) in parallel which is equivalent to reducing the search space for each one, and this now means there's no real benefit in restricting the velocity (Doppler shift) range, and the design is not made any easier by doing so.

 

Edited May 10, 2022 by IanD

[dixi188]

Coo!

It's a bit technical for us NB owners.

I used to use GPS and INS a lot when I flew for a living, but I never worried too much as to how the reults were derived, just used the answers.

Dixi.

[MtB]

2 minutes ago, dixi188 said:

Coo!

It's a bit technical for us NB owners.

 

 

 

I've been thinking that too.

 

I'm preferring the thread where we are discussing using a barge pole as a bow thruster.....

 

 

[IanD]

19 minutes ago, dixi188 said:

Coo!

It's a bit technical for us NB owners.

I used to use GPS and INS a lot when I flew for a living, but I never worried too much as to how the reults were derived, just used the answers.

Dixi.

One reason GPS has become ubiquitous is that it works much better then it used to, especially with initial locking speed. People don't realise how much it has changed "under the hood" since the first GPS receivers appeared, both in performance and cost -- technology advances are sometimes sadly unappreciated... 😞

19 minutes ago, MtB said:

 

I've been thinking that too.

 

I'm preferring the thread where we are discussing using a barge pole as a bow thruster.....

 

 

Ooh, old technology vs. new technology again -- I'll get some popcorn... 😉

[tree monkey]

To distract from the technical bickering and it's because of this thread I went off a googling to find the GPS unit that used to leave be stranded in the middle of fields for what seemed like hours whilst it attempted to find my position 

 

Anyway welcome  the technical wonder that was the thales mobile mapper, or more accurately "that bloody thing"

 

 

 

[furnessvale]

3 hours ago, MtB said:

 

 

I've been thinking that too.

 

I'm preferring the thread where we are discussing using a barge pole as a bow thruster.....

 

 

Indeed.  The last time I used a car based GPS to track a narrowboat down the River Weaver, it studiously placed the boat on the bank.  Somehow, it just could not accept the fact that the boat was out in the river.

[nicknorman]

8 hours ago, IanD said:

 

I explained that in my last post, but you obviously didn't catch it.

 

When trying to acquire satellites the receiver has to search through all the possible cases for delay and Doppler shift to lock on to each one. The bigger the speed range the bigger the Doppler shift, which means the search space is larger. If it's too large and the search takes too long then it can take a very long time or even fail to ever lock on.

 

This was what used to happen in the "old days" when processing power was limited by the processing speed in the chipset, and there was only one receiver searching for signals -- or at most a few, because more than this cost too much and consumed too much power -- so one "fix" (a bodge, really...) was to limit the Doppler search range by restricting maximum velocity.

 

Nowadays with modern chip technology the cost and power of each virtual receiver channel -- remember these are all digital signal processing, there's only one antenna input -- is absolutely tiny, transistor count has gone up by maybe 1000x since the design I was talking about, see attached plot. So they simply run a massive number (tens? hundreds? thousands?) in parallel which is equivalent to reducing the search space for each one, and this now means there's no real benefit in restricting the velocity (Doppler shift) range, and the design is not made any easier by doing so.

 

You are just paraphrasing what I said but in a way that gives you the last word. You hoped!