It sounds like an error in the model (or a bug) or perhaps an error in the
software you use to generate the test signal. Since the system works for zero
seperation, most everything is working correctly.
In article <XYydneTbNOYka3fZnZ2dnUVZ_t6dnZ2d@giganews.com>, "hmahdi12"
<hmahdi12@yahoo.com> wrote:

>Hello
>
>I designed a bistatic radar tracker with Extended Kalman Fileter. The
>model I that used is the constant velocity model. The measurements are:
>Doppler rang, Azimuth and Doppler frequency. the rang & Doppler
>measurements' covariance noise is low, but the Azimuth is too noisy. The
>state vector contained: x,vx,y,vy.
>My tracker works well when the distance between transmitter and receiver
>is zero (when the radar is monostatic). Otherwise, the estimation of
>azimuth has bias and the tracking performance is too bad!
>Is there any body can help me, please?
>
>
>
>

Reply by hmahdi12●August 22, 20062006-08-22

Hello
I designed a bistatic radar tracker with Extended Kalman Fileter. The
model I that used is the constant velocity model. The measurements are:
Doppler rang, Azimuth and Doppler frequency. the rang & Doppler
measurements' covariance noise is low, but the Azimuth is too noisy. The
state vector contained: x,vx,y,vy.
My tracker works well when the distance between transmitter and receiver
is zero (when the radar is monostatic). Otherwise, the estimation of
azimuth has bias and the tracking performance is too bad!
Is there any body can help me, please?