[Loadstone] better accuracy by position prediction

Shawn Kirkpatrick shawn at odyssey.cm.nu
Tue Oct 23 07:58:38 BST 2007

The position data from the receiver is usually only about a second behind, 
maybe less. The speed data is maybe 2 seconds behind. I don't think there's 
any way around this since you need some data to calculate speed. Since the 
speed and position are both being provided by the receiver I doubt you could 
use speed to correct position.
At some point there will be software calculations for heading and speed but 
that'll be to correct for those receivers that have proven almost useless 
for pedestrian use. Those receivers seem to give accurate position data 
thoe. If we had another source of data then maybe correction factors could 
be applied.

On Tue, 23 Oct 2007, Hasan Karahasan wrote:

> Hi folks,
> we all know that a gps based system takes some time to react on any changes.
> When you are standing, yourr speed will be zero at optimal gps conditions.
> When you suddenly start to move, the system needs several seconds to notice
> this and to offer the correct data. This happens to any change. The
> receiver's data is always some seconds behind the real data. Can we improve
> that indolent and sluggish behaviour by doing some prediction? I am sure
> that commercial software does similar things. We could do this by
> calculating an acceleration factor. While one thread measures the speed and
> calculates the accelleration periodically, another thread uses the result to
> correct the receiver's position permanently. If speed is in an increasing
> state, we know that the real position is more towards our current direction.
> We can add some meters to the receiver's position.
> If speed is decreasing, we can subtract something from the receiver's data.
> All loadstone functions would benefit from such a correction.
> What do you think about this?
> Hasan
> _______________________________________________
> Loadstone mailing list
> Loadstone at loadstone-gps.com
> http://www.loadstone-gps.com/mailman/listinfo/loadstone

More information about the Loadstone mailing list