car mph versus tomtom mph, why different?

droopsnoot

wongataa wrote: »

For live speeds GPS devices do use the doppler shifts of the satellite signals.

It seems to vary depending on the type of GPS receiver being used. On the one hand, technology has probably moved on since discussion such as this one - http://continuouswave.com/ubb/Forum6/HTML/002179.html - and this file linked from it - http://www.gps-results.com/GPS_Speed.pdf - which both date back more than five years. On the other hand, continual price reductions on consumer satellite nagivation systems will also force the individual components to be bought down to a price, so if a GPS receiver chip that calculates speed based on position comparisons rather than using doppler shift saves a few pence, I suspect it will be used given the small magnitude of the issue.

Richard53

Iceweasel wrote: »

Are we saying with 100% certainty that a Sat-Nav can only indicate the speed on a horizontal plane i.e parallel to the earth's surface?

I had always assumed this. The satellites 'see' a receiver moving across a plane parallel to the earth's surface (hence my Pythagoras calculation) and calculate the speed from that and their inbuilt time signal. Happy to be told otherwise, though.

bowlhead99

Richard, glad you're happy to be told otherwise, because, you are being...

As long as you can see 4 satellites you get a 3D position of where the receiver is relative to the rest of the globe. 3 satellites gets you a position on a flat map in two dimensions but adding a 4th gets you height as well. These days you can probably see 5-10 satellites depending on exactly where you are in the world and what trees and buildings are getting in the way..

So, your location data is in 3D. Even if the satnav receiver was literally plotting one position in 3D space and another position in 3D space a bit later and drawing a best-fit line between them and seeing how long it had taken itself to travel along that line: if the lines was a flat gradient between those points you are going to get close to the true distance and thus the true velocity.

Imagine you're in London atop the pointy 'Shard' skyscraper and you drive your go-kart down the side of it. Viewed from the moon, your journey from the top of the point to the bottom of one side might only seem to be 30 metres across, but in reality you've travelled 300 metres down hill, and if you know where you were in 3D space in the two separate instants, you know that you've gone a full 300m.

A key thing is that the measurement doesn't come from the satellites 'seeing' something from space and calculating where it is. It's not really like you or I sitting in a deckchair at a single point on the moon and saying, hmm, that looks like 30 metres across. The GPS receiver is the one doing the calculations. It sees 4 satellites in 3 dimensional space and from knowing where they are, it knows where it is at the start of its journey and the end of its journey. So it could calculate the 300m and thus its speed, if it wanted to.

However, for a short distance, say just half way down the tower, the fact that the first position has a margin of error of 10m and the second position has a margin of error of 10m, you are looking at 20m of error in less than 200m. And if you were snaking down the side of the building rather than going in a dead straight line, you would be covering quite a different distance anyway - whether you were doing this down the side of a skyscraper or great pyramid or just a chicane on a race track or the curves of the River Thames. For this reason they basically don't just take two snapshots and see how long the line joining the dots theoretically might be.

In reality they know you want, every fraction of a second, an instant speed readout. They can and do work this out from 'listening' to the doppler shift. As you zoom along a plane away from a satellite, the beeps or pings coming from it get further apart. As you go towards a satellite they appear to get closer together. When you have 3 or 4 of them you can put the data together and see how fast you're going. You don't need to see where you were and where you are now and how long it took and guess what distance it was and what speed was implied.

Some further reading http://books.google.co.uk/books?id=6P7UNphJ1z8C&pg=PA92&redir_esc=y#v=onepage&q&f=false if you're interested

Measurement of the received carrier frequency provides information that can be used to determine the velocity vector of the user. Although this could be done by forming differences of code-based position estimates, frequency measurement is inherently much more accurate and has faster response time in the presences of user dynamics...

If you read 3.5.5 and 3.5.6 it should help you get a handle on it ; perhaps best to just read some of the words rather than trying to work through the formulae which looks a little more intimidating than Pythagoras...
:beer:

IamNotAllowedToUseMyName

Richard53 wrote: »

I had always assumed this. The satellites 'see' a receiver moving across a plane parallel to the earth's surface (hence my Pythagoras calculation) and calculate the speed from that and their inbuilt time signal. Happy to be told otherwise, though.

You aren't entirely wrong because GPS systems work in a number of different ways. If you have used a Garmin Edge as we cyclists use, you will also find that the GPS system is useless for height as it uses a theoretical model of the earth (as the earth is not round so the distance to the surface from the satellite is not consistent), so we have a barometric altimeter to get a far better height (reasonably accurate relative to a starting height). On the Garmin site, you can compare a GPS mapped altitude profile of a ride with a map data calculated one - the differences are massive.

I would suggest that the majority of car satellite systems ignore any attempt at calculating height as it is too complex to no benefit. Systems that do use height, such as the Mercedes automatic gearbox system for its top of the range lorries, will acquire their height from surveyed height data stored in the mapping and use the GPS position to look up the height from the mapping data, not from using GPS (and of course it is looking ahead along the route to see what the profile of the route is which obviously cannot be driven by getting GPS data for half a mile further down the road).

I have an older Garmin eTrex Legend that calculates height using GPS and at times it quite happily tells me when I am 15 metres below sea level. This is not a fault of the electronics, it is a limitation of the mathematical model of the earth that is used.

br3nnan

IanMSpencer wrote: »

On the Garmin site, you can compare a GPS mapped altitude profile of a ride with a map data calculated one - the differences are massive.

Completely off topic but what is the difference you get. Is it in the 47-50m range.

wongataa

IanMSpencer wrote: »

You aren't entirely wrong because GPS systems work in a number of different ways. If you have used a Garmin Edge as we cyclists use, you will also find that the GPS system is useless for height as it uses a theoretical model of the earth (as the earth is not round so the distance to the surface from the satellite is not consistent),

The geoid used in the WGS84 map datum (used by GPS) doesn't exactly match the real planet which can account for some elevation errors, but the height component of GPS data is less accurate than the horizontal data due to how the system works.

benjus

If it's using doppler shift to calculate the speed it's irrelevant what kind of geoid the position is projected onto.