Hi! We see you’re using an ad-blocker. We’re fine with that and won’t stop you visiting the site.
But as we’re losing ad-revenue from this then why not make a donation towards website running costs?. Or you could disable your ad-blocker for this site. We think you’ll find our adverts are not overbearing!
Joined: 11/03/2003 08:29:43 Posts: 64 Location: United Kingdom
Posted: Fri Mar 24, 2006 3:55 pm Post subject:
Lost Cause wrote:
Hmmm… well again that is a customer service centre response (whose job it is to answer as many enquiries as quickly as possible – and is not manned by engineers).
Our customer service IS manned by engineers, whose job is to assist customers in integrating our GPS receiver modules into their applications. Half of the TDC staff are engineers.
Joined: 11/03/2003 08:29:43 Posts: 64 Location: United Kingdom
Posted: Fri Mar 24, 2006 4:10 pm Post subject:
And compensating for doppler would be quite complex as the satellites are themselves around you at a reasonable speed! (Although the receivers do compensate for time and gravity dilation which is pretty amazing stuff! 8O ) _________________ G1NTX AFRIN LRPS
Thanks for not calling me an engineer! The short answer was because I was busy, but I rise to the challenge -
Simon
Lol ! Nice one !
Ha, maybe I jumped to that conclusion because if you asked any of the distributors of our products an equivalently technical question then the chances of getting an insightful answer would be zero ! The quality of staff at distributors can vary enourmously, not least those seeking to answer customer queries. So don't take any sleight in this particular instance ! Therefore, TDC is a distributor in the UK for a variety of GPS chipsets/components, correct ?
I'm a bit confused because there is a great deal of reference on the net (with a bit of Googling) to a variety of cheap and expensive GPS receivers using Doppler shift/carrier phase to calculate velocity. In fact, the more expensive solutions (such as Vbox) seem to rely upon it - thus indicating that attempts to measure speed in this manner are not futile. Earlier in the thread TomDavison was roundly rejecting the principle outright however, based upon his experience in the military with radars. Clearly there is more than one way to skin a cat. Me, I've got no idea how they do it, I'm trying to find out.
Being the UK distributor of Sirf chipsets, then I wonder if you could use your influence to ask one of the DSP guys at Sirf as to how they do it in their particular case ? Seems how Sirf is a popular solution in a lot of low-end products (such as my wonderful Fortuna device). Assuming you've all got nothing else better to do of course.
Thanks for not calling me an engineer! The short answer was because I was busy, but I rise to the challenge -
Simon
I'm a bit confused because there is a great deal of reference on the net (with a bit of Googling) to a variety of cheap and expensive GPS receivers using Doppler shift/carrier phase to calculate velocity. In fact, the more expensive solutions (such as Vbox) seem to rely upon it - thus indicating that attempts to measure speed in this manner are not futile. Earlier in the thread TomDavison was roundly rejecting the principle outright however, based upon his experience in the military with radars. Clearly there is more than one way to skin a cat. Me, I've got no idea how they do it, I'm trying to find out.
Lets go back to basics. Doppler speed measurement using radar is a standard, accurate procedure. If you are going directly away from or towards the transmitter it is straightforward. If you are going at an angle to the transmitter, but you know your precise location and direction, you can do the maths to correct the recorded speed to the actual speed. If you don't have a position and direction of travel, the doppler shift could be identical for two targets, one moving slowly towards or away from the transmitter, and one moving faster, but crossing diagonally.
If the transmitter is also moving, you have to factor in the doppler shift that is caused by the movement of the transmitter, which is how AWACs track movement . The AWACs knows its own location, track and speed and can therefore factor out its own doppler effect.
With satnav, the instantaneous doppler shift of a satellite will vary for any receiver as the satellite is approaching, or receding. Unfortunately, since the satellite is not on the same plane, the apparent speed towards the receiver varies continually, being highest as the satellite comes into view over the horizon, reducing to a minimum as the satellite passes a line drawn through the receiver at 90 degrees to the satellite travel, and then increasing again as the satellite heads off towards the horizon. The closer the satellite is to passing directly overhead, the bigger the difference.
If you know your accurate position, and direction of travel, and the instantaneous doppler shift that would be obtained from a satellite if you were stationary, it is theoretically possible to measure the actual doppler shift, and determine the element of that shift that can be attributed to target motion. (The transit system used by the USN did a variation of this. They used known direction of travel and speed to calculate position, from a single satellite)
If you already know your accurate position, and direction of travel, it's simpler to just do the maths between known positions. Problem with this is that you are calculating average speed over the time period. If you want to measure acceleration/deceleration accurately, this doesn't work. Accelerate or decelerate hard, and see how far behind your satnav lags. But for practical purposes, if you accept that speed is only accurate if you are travelling straight and level without acceleration/deceleration, what the hell, satnav speed is good enough and still more accurate than vehicle speedos.
Without having to calculate satellite induced doppler shift, life becomes much more simple and I can see how the use of a static base station might permit doppler solutions. If you know the instantaneous shift caused by the motion of each satellite, (cos you are measuring it at the base station), any variation to that measured shift on a receiver nearby must be caused by receiver motion. Take into account sufficient satellites and it must be possible to identify a unique velocity which would account for those shifts. This might permit very accurate speeds, including acceleration/deceleration to be obtained, close to the base station. (perhaps on a racetrack, proving ground etc) What I can't see is how any system can sensibly do this without the base station, or at any distance from the base station.
So I go back to my premise that for satnav, use of doppler solutions is a non-starter.
Joined: 06/06/2003 10:17:21 Posts: 57 Location: United Kingdom
Posted: Fri Mar 24, 2006 6:47 pm Post subject:
TomDavison wrote:
Without having to calculate satellite induced doppler shift, life becomes much more simple and I can see how the use of a static base station might permit doppler solutions. If you know the instantaneous shift caused by the motion of each satellite, (cos you are measuring it at the base station), any variation to that measured shift on a receiver nearby must be caused by receiver motion. Take into account sufficient satellites and it must be possible to identify a unique velocity which would account for those shifts. This might permit very accurate speeds, including acceleration/deceleration to be obtained, close to the base station. (perhaps on a racetrack, proving ground etc) What I can't see is how any system can sensibly do this without the base station, or at any distance from the base station.
Well I think there is some convergence of ideas going on here - but what remains different ? The problem of measuring satellite induced doppler shift.
From my reading, it's the modelling of the system that saves the day here. The designers know the motion, speed and altitude of the satellites i.e. their orbital characteristics. Once the receiver's position is known, and the position and velocity of the satellite is known (as transmitted conveniently by the satellite) then so an algorithm can be written up to calculate the Doppler shift (with respect to a stationary receiver at any point on the earth). If the satellites were in random motion then the situation would become impossible - but they behave in a predictable manner. A manner that allows for an algorithm to be calculated for different relative positions in sky and on planet, of receiver and satellite. As the position and trajectory of the satellite is broadcast in each case, so you only need to find your position - and then plug that into the algorithm. Out pops a calculation of the Doppler shift (relevant for a static receiver) for any point on the planet, for any satellite found at any particular position. This is what I've read from a number of sources. At this point you can do the latter part of what you describe (analysing actual measured Doppler shift, at that same point in time, in each channel to get a vector.)
As for delays, then that is certainly going to happen on a cheaper receiver sampling at only 1Hz. Plus, are you reading the figure given by your routing software, or the speed given by the GPS itself ? I'm convinced that TomTom would smooth any speed measurements further (averaging it out over time) causing further delays. Why ? Well as mentioned previously, to stop the speed readout become erratic in poorer reception conditions. In the same way as they might perform the same thing on position data. So I can see good reason for significant delays in the routing software, because in that instance it is actually performing a longer sample. Whereas if you install an application on your PDA that allows you to monitor the raw data coming off the GPS directly, then you would in all likelihood see a lesser delay. And note that the more expensive GPS solutions (Vbox) go to sample every 20th of a sec instead - and in that case does even a good GPS receiver have sufficient accuracy on position information to allow super accurate speed measurements on a distance vs time calculation ? At a steady 40mph, consider the ground being covered by the car every second; can Vbox really track motion (in terms of car position) with GPS that accurately at 20Hz and still give a super accurate speed readout all the way between each successive 1/20th sec sample ?? It certain seems able to. Yet Vbox doesn't have the luxury of doing long averaging time - it's chasing the 0-60mph times of Ferraris after all !
Anyhow, it's all interesting stuff (hey otherwise why the long posts !). Given that none of us actually 100% know the answer to this (not being programmers of these things), then no amount of debating amongst ourselves will resolve the issue. So it might be worth you taking a quick timeout and type something like 'Doppler Gps speed calculation' into Google and reading through some of the hits - because that's one of the things I've done. Then see what they've got to say, because to be honest, it's only their stuff I'm repeating - and the stuff I'm reading seems to make sense. They might all be wrong, of course, but it would be worth reading the stuff of those more knowledgeable directly, rather than me trying to paraphrase it for them. If you see what I mean. Then come back and we'll continue the debate for the answer !
VBOX -128K RAM
GPS magnetic aerial
12 hour battery pack with built in 12v charger
Mains charger specify UK, Europe, Switzerland, Italy, Denmark, North America or Australia
Cigar lighter adapter
CD ROM containing VBOX software
Serial PC cable for download / real time monitoring
User manual
RLVBOXPROF VBOX Pro £4,995
As standard VBOX specifications plus
1Mbit CAN link for integration into existing Data logging equipment
1Mbyte battery backed RAM
Lateral / longitudinal g sensing
Lap timing function
1 x Padded carrying case
RLVBOXMOT VBOX Motorsport £4,995
As standard VBOX specifications plus
1Mbit CAN link for integration into existing Data logging equipment
1Mbyte battery backed RAM
Lightweight housing
Waterproof Deutsch Autosport connectors
Spares
VBOXAERIAL GPS Aerial £75
RLVBOXBATT2 Battery Pack 12hr with built in 12v charger £150
RLVBOXCHG2 Mains charger specify UK, Europe, Switzerland, Italy, Denmark,
North America or Australia £65
VBOXCIGADAP Cigar lighter adapter £35
Accessories
RLVBOXDIGITA LED speed readout display (dashboard or windscreen mount) £325
RLVBOXRAD2 Telemetry link (10 mile line-of-site range) 2 x Radios with cables £1950
RLVBOX4ANIF 4 channel Analogue interface, 1 analogue output (VBOX pro only) £350
VBOXBRAKTR Brake pedal trigger £75
Joined: 06/06/2003 10:17:21 Posts: 57 Location: United Kingdom
Posted: Sun Mar 26, 2006 3:06 am Post subject:
What`s your thoughts on whether Vbox is working on a GPS Doppler system or not ? Their website says it is.
For sure Vbox is a high-end solution, but then it is intended to be a reference system. Doesn`t mean to say low-end solutions are not using the same principle.
I have a 10 quid astronomy program running on my PDA. If I input my lat/long and the time of day then, without any measurement, my PDA will tell me the exact position and trajectory of the International Space Station and other space vehicles. I sometimes use it to observe Iridium flares. It will tell me how far above the horizon etc, etc everything can be obsevered at. It`s able to do all this (for a tenner) because it`s a mathematical problem - not a measurements exercise (beyond knowing some simple things like current time and position - which GPS does very well). It`s very predictable - unlike the motion of military jets. In fact, the people writing the algorithms can go wild if they want to and include the effects of relativity, the atmosphere, anything they can think of basically, to feed in to improve the accuracy of the prediction. As it`s purely a maths thing, not a measurement thing, so all the astronomy programs allowing you to track satellites and predict their positions are quite cheap. Even the low volume PDA ones.
Therefore, knowing where these things are in the sky, relative to you, for any time of day is not tricky at all, as my PDA application proves. With the trajectories of the GPS sats well known also (and updated with their own trajectory transmissions too), then predicting Doppler induced satellite shift is purely a case of calculation. With this highly accurate frequency figured predicted, you can compare it with your actual reading and find out your velocity. By taking advantage of the atomic clock accuracy (with enough sats locked onto) within my cheap GPS receiver, this allows for a) me knowing exactly what the time to feed into the algorithmis and b) to make an excellent frequency/phase measurement of the sats. I`m now in fine shape to compare actual measurements with predicted meaasurements. Both of which are very precise figures. Hence I can claim 0.1m/s accuracy on my 100 quid GPS receiver.
Now, the GPS receiver is going to be using all this astronomy type programming already, as part of its positioning task.
Then, of course, if you want to go out and drop a small fortune on a reference system (for your motoring magazine) that operates at 20Hz rather than 1Hz, then you can buy a Vbox. Actually, I thought they would be dearer than that.
Note all of this would be entirely useless, in your job, for tracking an F16. Because an F16 flies in an intentionally unpredictable way in combat. But as my 10 quid astronomy app proves, knowing the whereabouts and direction of a satellite above you, ten days from now, six months from now - relative to my own position - is child`s play in comparison. And with enough effort at the programming stage, you can know that figure really accurately. At that point we just need a really accurate frequency(or phase) reading from each sat for comparison, courtesy of our terrific atomic clock quality timepiece on-board.
Joined: Mar 02, 2006 Posts: 384 Location: Bedford, England
Posted: Sun Mar 26, 2006 9:29 am Post subject:
Lost Cause wrote:
What`s your thoughts on whether Vbox is working on a GPS Doppler system or not ? Their website says it is.
. At that point we just need a really accurate frequency(or phase) reading from each sat for comparison, courtesy of our terrific atomic clock quality timepiece on-board.
That theory seems quite appealing to me.
The theory is appealing, certainly, but we do need to measure very accurately the received frequencies of the various satellites. Do we have the technology to do that within the price/capabilty of our vehicle satnavs.
Certainly, when I was working with the radars we didn't have the ability to do so. Admittedly, we were only working with a single frequency source, and not multiple sources. I am prepared to admit I am wrong, IF I can be persuaded that the technology exists within the chipset which forms the GPS element of our satnav to determine location by differential times of transmission, calculate satellite doppler shifts for a stationary object at that location, measure actual doppler shifts for the various satellites, filter out the shift applicable to satellite motiion, and compute a receiver velocity to account for the remaining shifts.
As you say, most of it is maths. What I can't see is the abilility to measure multiple frequencies simultaneously, to the degrees of accuracy that will allow you to filter out a satellite induced shift of thousands of miles per hour, and still be able to read a receiver induced shift of one or two miles per hour. And do it in a chipset that costs a fraction of the total cost of the couple of hundred pounds of the satnav.
Joined: 06/06/2003 10:17:21 Posts: 57 Location: United Kingdom
Posted: Tue Mar 28, 2006 3:55 pm Post subject:
TomDavison wrote:
The theory is appealing, certainly, but we do need to measure very accurately the received frequencies of the various satellites. Do we have the technology to do that within the price/capabilty of our vehicle satnavs.
Well that's a very good question on which the whole thing hinges. And never having built such a device, then I've no idea of the answer. Because it's a maths exercise, then I've no problem with the whole satellites themselves moving thing, but the GPS receiver being able to measure these frequencies accurately is vital to the success of the idea. My guess has been that the good clock on-board is a step in the right direction, but whether that gets us all the way there is another matter. I wonder if one day we'll find out for sure ?
The theory is appealing, certainly, but we do need to measure very accurately the received frequencies of the various satellites. Do we have the technology to do that within the price/capabilty of our vehicle satnavs.
Well that's a very good question on which the whole thing hinges. And never having built such a device, then I've no idea of the answer. Because it's a maths exercise, then I've no problem with the whole satellites themselves moving thing, but the GPS receiver being able to measure these frequencies accurately is vital to the success of the idea. My guess has been that the good clock on-board is a step in the right direction, but whether that gets us all the way there is another matter. I wonder if one day we'll find out for sure ?
I'm not convinced you need to 'measure' very accurately the received frequencies, in the way implied. What you need is to be able to *vary* the receiver frequency accuracy. Then, knowing pretty much what it will be within very tightly defined limits (you know the speed of the satellite and the earth, so the only variable is the receiver), you can vary the reception frequency until you have tuned in.
All the other stuff,as was very cogently pointed out, 'is just maths' and requires no special hardware all all.
So, how much does a very accurate frequency variable receiver cost? Not much, or GPS chipsets (with all the other gubbins) wouldnt be available for well under $50, and we know that a GPS unit *must* have a very accurate frequency adjustable receiver ***or it wouldn't work at all in the first place***.
Joined: 24/06/2003 00:22:12 Posts: 2946 Location: Escaped to the Antipodies! 36.83°S 174.75°E
Posted: Tue Mar 28, 2006 11:08 pm Post subject:
TomDavison wrote:
As you say, most of it is maths. What I can't see is the abilility to measure multiple frequencies simultaneously, to the degrees of accuracy that will allow you to filter out a satellite induced shift of thousands of miles per hour, and still be able to read a receiver induced shift of one or two miles per hour. And do it in a chipset that costs a fraction of the total cost of the couple of hundred pounds of the satnav.
I see your point, it does sound like an incredible feat of engineering but as tumbleweed points out, the whole concept of GPS positioning is mind boggling.
Even when stationary, the GPSr manages to track 12 satellites and triangulate your position by measuring the distance to 4 satellites, accurate to within about 30 feet. That's pretty incredible when the sats are over 12,000 miles away and moving at thousands of Miles per hour.
And all this in a receiver that costs £50. Nah, it doesn't sound possible, but as we know it's true! 8O
Now add a bit of doppler shift? That would be a doddle. _________________ Gone fishing!
Joined: Mar 02, 2006 Posts: 384 Location: Bedford, England
Posted: Wed Mar 29, 2006 1:24 am Post subject:
I have to confess that I am starting to believe in doppler shift, even though when I was in the RAF the science did not exist to achieve what must be happening if this is the answer to velocity measurement.
Theoretically, I can see no problem to using doppler shift. There are some seriously BIG ifs, however.
The options are
1. First you need to know your accurate position. Then a finite period later, you need to know a second accurate position, and the time between them. It's then simple maths to calculate the position.
2. You need to know your accurate position. Then you need to know the accurate position and velocity of the satellites you are receiving. This will allow you to calculate the rate of approach or departure of the satellite from your location. Based on this, you calculated the doppler shift that you would expect from each satellite if you were stationary. Then you measure the frequency of each satelllite and determine the actual doppler shift. From this you derive a velocity which accounts for the difference.
Problems are: If the satellite is close to the vertical, there is little or no doppler shift. If your location is at 90 degrees to the satellite's position and track, there is no doppler shift. The doppler shift from the satellite is never steady, and is always changing at a rate which varies with the distance of the satellite from you, and the direction of travel of the satellite. The doppler shift also varies with your direction, and turning a corner can change from a recordable shift to no shift. Finally satnav can give speeds of 1 or 2 mph. If you are looking at a satellite which is travelling at 2000 mph, and you are travelling at 1mph it needs to identify a doppler shift from the vehicle motion which could be around .0005 of the total shift. Can we really measuer frequencies from multipel satellites to this order in the chipset?
If we can, I will eat humble pie, and accept that this is how it works. Convince me.
I have to confess that I am starting to believe in doppler shift, even though when I was in the RAF the science did not exist to achieve what must be happening if this is the answer to velocity measurement.
Theoretically, I can see no problem to using doppler shift. There are some seriously BIG ifs, however.
The options are
1. First you need to know your accurate position. Then a finite period later, you need to know a second accurate position, and the time between them. It's then simple maths to calculate the position.
2. You need to know your accurate position. Then you need to know the accurate position and velocity of the satellites you are receiving. This will allow you to calculate the rate of approach or departure of the satellite from your location. Based on this, you calculated the doppler shift that you would expect from each satellite if you were stationary. Then you measure the frequency of each satelllite and determine the actual doppler shift. From this you derive a velocity which accounts for the difference.
Problems are: If the satellite is close to the vertical, there is little or no doppler shift. If your location is at 90 degrees to the satellite's position and track, there is no doppler shift. The doppler shift from the satellite is never steady, and is always changing at a rate which varies with the distance of the satellite from you, and the direction of travel of the satellite. The doppler shift also varies with your direction, and turning a corner can change from a recordable shift to no shift. Finally satnav can give speeds of 1 or 2 mph. If you are looking at a satellite which is travelling at 2000 mph, and you are travelling at 1mph it needs to identify a doppler shift from the vehicle motion which could be around .0005 of the total shift. Can we really measuer frequencies from multipel satellites to this order in the chipset?
If we can, I will eat humble pie, and accept that this is how it works. Convince me.
Tom
OK lets reverse engineer this.
60mph is roughly 25 metres a sec.
Natural discrepancies in your position can easily be 10m in a second.
For example, if you have a GPS unit that doesnt do smoothing, such as some hand held outdoors units, you will see your actual position move about by about 10 metres. Lets take 10m as being average.
If you did the position / time calculation as per your argument, your position at the start of a second might be +/-10m away from your true position. Your position at the end of that second might be +/-10m. So, your measured position difference, when you have actually done 25m, might be anywhere from 5m to 45metres. So if you updated the speed every second, which I think many car units do, the speed would fluctuate wildly. In fact, it would fluctuate wildly *even if you weren't moving*, and even if a longer term (say rolling over past 10 seconds)smoothing algorithm was applied. (which would be really crap for in car use) Even if stationary you'd expect to see your speed sometimes reported as say 10mph, just due to random fluctuations, or to dip lower for a second or two as you were accelerating, which it never does.
Whereas this sort of thing *must* happen if the TT measures your position and then uses that to do the calculation, even if it locks you to a road (which it does), because it wont be able to fully damp out the random position fluctuations, because it doenst know they are random fluctaions, maybe you really did move to that new position?
Convinced? If not, switch your TT on, put it on the windowsill,and wonder why it doesn't sometimes report your house moving at 20mph!
Please correct me if I'm wrong but didn't an engineer from the chipset supplier state that doppler shift wasns't used? He said that the location could be up to 10m out but that subsequent locations could be very accurately measured in relation to the original location which would give an accurate speed between the two.
He also mentioned that some sat nav systems would dampen the speed output to avoid variations in speed when the sats being used changed as this may introduce a further 10m error.
My RoadPilot Micro often reports speeds of upto 4mph briefly when stationary due to, presumably, the speed output not being dampended.
Joined: 24/06/2003 00:22:12 Posts: 2946 Location: Escaped to the Antipodies! 36.83°S 174.75°E
Posted: Thu Mar 30, 2006 1:45 pm Post subject:
Perhaps someone could try logging the raw NMEA data stream and then measure the difference between the two reported positions and compare that to the speed that that is reported in the NMEA data. _________________ Gone fishing!
Posted: Today Post subject: Pocket GPS Advertising
We see you’re using an ad-blocker. We’re fine with that and won’t stop you visiting the site.
Have you considered making a donation towards website running costs?. Or you could disable your ad-blocker for this site. We think you’ll find our adverts are not overbearing!
You cannot post new topics in this forum You cannot reply to topics in this forum You cannot edit your posts in this forum You cannot delete your posts in this forum You cannot vote in polls in this forum
Or you could disable your ad-blocker for this site. We think you’ll find our adverts are not overbearing!
Hi! We see you’re using an ad-blocker. We’re fine with that and won’t stop you visiting the site.
But as we’re losing ad-revenue from this then why not make a donation towards website running costs?. Or you could disable your ad-blocker for this site. We think you’ll find our adverts are not overbearing!
Forum FAQ
Search
Usergroups
Profile
Log in for private messages
Log in
