[Loadstone] Direction change
rob at loadstone-gps.com
Tue Dec 2 07:39:52 GMT 2008
Nothing to add. that's how GPS works, but some have a hard time believing it.
At 12/2/2008, you wrote:
>Gps direction is rarely accurate as it can only calculate a change in
>direction based on previous locational data.
>Given that the update interval on most receivers is 1 second, and by
>definition short-term direction changes are inaccurate most software
>introduces a fudge factor which says "don't change the direction unless you
>see 3 seconds of this"
>Also note that the direction is undefined when you are turning as the area
>involved is small (smaller than the accuracy of gps in many cases) and your
>direction is changing with little change to location.
>Try this for an experiment:
>Stand still, hold the gps receiver the right way around. now turn the
>receiver by 180 degrees. Note your direction does not change.
>This shows that it is the co-ordinates received by the receiver that give it
>its idea of direction and not the orientation of the receiver. This means
>it'll give accurate readings when it is upside down in your pocket.
>Now take the readings from a gps with static navigation turned off and the
>direction will bounce around from second to second, north, east, south,
>north, west, southwest and so on.
>A gps needs a history of readings to calculate direction, and by default the
>update interval is 1hz. As discussed on this list, increase that and you'll
>potentially overload the phone with data it might not be able to use, and
>the processing requirements increase potentially beyond the ability of the
>phone to cope.
>It also depends on the firmware of the gps as to which nmea sentences it
>sends when; there is a program with gpsd for Linux that will show you how
>often these sentences get sent. Not sure if SiRFdemo can do this also.
>If loadstone doesn't get timely data from the gps, it can't use it. If it
>gets bad data, it can't use it.
>It would be well to remember you are using a series of satelites in
>geostationary orbit to track your location; and not a differential source.
>If you had a ground-based differential source and the ability to use it;
>you'd be able to potentially get far more responsive direction information.
>In short you are asking potentially too much of a program that runs on a
>slow hardware platform with unreliable to downright useless gps information
>from a given receiver in a potentially hostile urban environment depending
>on where you live.
>You have multi-path echo, drift and interference to contend with.
>Shane Rob and co feel free to correct any shortcomings in this explanation
>as i'm doing it from memory.
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the Loadstone