## INS / GPS data fusion using Kalman filter - need help for tuning

### INS / GPS data fusion using Kalman filter - need help for tuning

Hi there,
I'm working on a low cost integrated INS/GPS navigation system (fo
automotive applications).I use only 1 2D accelerometer and a GP
receiver.
This is my system model (2D movement):
x(k) = Ax(k) + Bu(k) + w
z(k) = Hx(k) + v

Where:
x(k)=[sx sy vx vy] (sx = position along x axis, sy = position along
axis, vx = velocity along x axis, vy = velocity along y axis).

u(k)=[ax ay] (ax = acceleration along x axis, ay = acceleration along
axis), this is the output of the gyros.

z(k)= GPS measurements (position along x and y).

A=[1 0 T 0; 0 1 0 T; 0 0 1 0; 0 0 0 1];
B=[T^2/2 0; 0 T^2/2; T 0; 0 T];
H=[1 0 0 0; 0 1 0 0];

T= sampling time;

I've some problems in filter tuning, do you have some suggestions on ho
can I evaluate Q and R matrix?

Thanks and sorry for my english ^_^

### INS / GPS data fusion using Kalman filter - need help for tuning

You might also wish to ask on sci.geo.satellite-nav

This would be the place for the math.

sci.geo.satellite-nav has some developers who read the group. It has a
strong end user emphasis now, but it's original intent was to serve
developers.

I know little of either field, but am fascinated by both.

BTW, your English is fine. I know many high school and college
instructors that wish their native English speaking students were as fluent.

For perspective, ~50 years ago while vacationing in northern Ontario,
Canada [a predominately French speaking area], our host asked me to
borrow something from a neighbor. I tried out my French. I was asked to
speak English as it was much more recognizable than my French ;\