[Spridgets] GPS

[Guy R Day]

Speed is a function of time AND distance.

You cannot accept the time is accurate - as that relates to the variance in 
signal distance - although I accept if you are locking onto one satellite 
that production of that time signal will be a constant, (although the 
reception of it may not) you still have the error with the time factor in 
the satellite distance component.  The processors do not anticipate and 
correct, they deal with the data provided, not imagined or predicted data.

0-60mph of 8.3s suggests a distance of shortly over 365ft (111m) and around 
1/3rd g acceleration.  Quite respectable, no problem with that (remove the 
spare wheel, take a seat out and dispose of other junk in the trunk etc and 
improve on the figure <smile>).

Now start messing with the figures provided and the distance can be 109 to 
113m if you are within the 2m error distance.  You may well be outside the 
error factor as that is a mean, so let us say the variance is really 4m, 
that is 107 to 115m.  You have inaccuracy built in to the device by the very 
nature of its limitations.  The signal you receive varies with the motion of 
the vehicle and the reception of the signal.  It is not just the strength of 
the signal but the reflections and bounces of it as well that are used and 
in a moving car etc etc.

You know your ordinary speedo in the car shows a definitive figure but you 
also know that is not accurate.  The figure produced by the device (speedo 
or gps setup) may show a definitive figure with fractions of a second but 
that does not mean it is an accurate figure.  The accuracy falls under the 
maxim of 'you can fool some of the people most of the time etc....' <smile> 
Some of the GPS devices tested for use in skid testing would not even show 
it being used on the right road (about 60m away) and the manufacturers 
supplied them knowing they were to be used for accurate measurement in a 
series of 30-0 and 40-0mph skid testing sessions.

For a figure used by the guy in the street they are as good as anything else 
they use.  For an accurate figure they cannot be trusted.  Check it against 
a Vericom 4000 or similar - but make sure that device is nailed in the 
correct position within the car and not just slung on the dash - yet another 
source of inaccuracy!

The question was, 'is GPS accurate?'  The answer is no.  Had the question 
been, 'is a GPS a useable bit of kit suitable for use by the man in the 
street?', the answer is yes.

There is, of course one other problem - what happens when the GPS signal is 
lost altogether?  What is the signal like in the high rise areas of your 
cities - or down in a tunnel or cwm? <smile>?


Guy R Day



----- Original Message ----- 
From: "Ron Soave"
To: "Bugeye" ; "Spridget" ; "Larry & Sandi Miller" ; "Guy R Day" Sent: 
Monday, September 20, 2010 4:37 AM
Subject: Re: [Spridgets] GPS


--- On Sun, 9/19/10, Guy R Day  wrote:
> Ron, Explain why
> please.

Most are accurate to 2m or so, the time part is dead nuts, processors are 
plenty quick enough to anticipate and correct, and I think that the speeds 
of a 0 - 60 run, that's plenty accurate. This week many of the cars at Targa 
Newfoundland were using GPS as a supplement to their Terratrip rally 
computers.

Ron