kalman filter based azimuth and elevation tracking models

does anyone knows literature concerning models for kalman filter based
azimuth and elevation tracking(without using range).Is it necessary to go
to cartesian system of coordinates for kalman filter model?

On Jul 11, 12:50 pm, "Sylvia" <sylvia.za...@gmail.com> wrote:
> does anyone knows literature concerning models for kalman filter based
> azimuth and elevation tracking(without using range).Is it necessary to go
> to cartesian system of coordinates for kalman filter model?

I think there are several options, Cartesian is often chosen because
of the ease of the target motion model, but it often leads to strong
nonlinearities in the measurement model. For cases where the range is
unobservable the modified polar/spherical is popular, since the
unobservable component is limited to one state, and it becomes
observable when an acceleration takes place. It still requires and
Extended Kalman Filter or some other mechanism to handle the non-
linearities.

Blackman's book "Design and Analysis of Modern Tracking Systems"
discusses this and provides quite a number of references.

Cheers,
David

Sylvia wrote:
> does anyone knows literature concerning models for kalman filter based
> azimuth and elevation tracking(without using range).Is it necessary to go
> to cartesian system of coordinates for kalman filter model?

Are you tracking orbiting satellites for which you know the elements? If
so, there are quite a few programs around that will drive antennas for
you, or at least give you the azimuth/elevation figures. Many will also
compute Doppler shift.

See http://www.amsat-na.com/category.php?c=Software

I think there are also some GPLed packages for Linux/UNIX.

Most satellite tracking programs take Keplerian orbital elements, but a
few take cartesian state vectors, and there are utilities that will
convert between them for you.

--Phil