I need help implementing Extended Kalman Filter

Dear everyone,

I need to implement the Simulink Extended Kalman Filter block but I don't know how.

I have the simulink scheme

this is the system code inside the matlab function

A = [0 1 0 0;

0 -2*(Cf+Cf)/(m*V) 2*(Cf+Cf)/m -2*(Cf*lf-Cf*lr)/(m*V);

0 0 0 1;

0 -2*(Cf*lf-Cf*lr)/(Iz*V) 2*(Cf*lf-Cf*lr)/Iz -2*(Cf*lf^2+Cf*lr^2)/( Iz*V)];

B1 = [0 2*Cf/m 0 2*Cf*lf/Iz]';

B2 = [0 -2*(Cf*lf-Cf*lr)/(m*V)-V 0 -2*(Cf*lf^2+Cf*lr^2)/(Iz*V)]';

C = [1 0 0.7 0];

xdot = A*x + B1*u + B2*psi;

y=C*x

the state is not measurable and at the input the V is not constant, but a sine, I should implement an Extended Kalman Filter in Simulink to estimate the state of this system.

Could anyone help me how to do it?

[Sanju]

can you please share your model for better understanding.

[Umberto Tozzi]

@Sanju of course, this is my model. Could you help me implement Extended Kalman Filter in Simulink to estimate system state?

The model only gives me the y. The x in question is not measurable.

Answer 1

[Muskan]

Hello Umberto,

The formulation of an extended Kalman is based on the linearization of the state equation and measurement equation. Here is the officicial documentation on how to implement Extended Kalman Filter in MATLAB: https://in.mathworks.com/help/driving/ug/extended-kalman-filters.html

I hope this helps!