A question about extended kalman filter block

Hi everybody. I have a problem in implementing extended kalman filter in simulink for a quarter car active suspension system. Has anyone worked with the ekf block so he can help with my problem? And I explain it in more details?

Actually when I use the block as an open loop without controller, it estimates states properly. But when I add the controller u to the block, it doesnt work!

[praguna manvi]

Hi @Faezeh Yousefpour,

Can you share your model for the same.

[Pavl M.]

Good question desires entire settlement, full working solution-answer.

How do you interconnect State Estimator with Controller and Plant?

Which controller kind is it?

What is the rationale and motivation for EKF versus Particle Filter or ExtendedRunningAverage?

Which noise and disturbance how modeled, how handled?

What for review and novel Research within Controller integrated(united) in Sensor scope?

What for scripted interconnection (vs block diagram) between EKF, controller and plant?

Whether EKF auxilliary matrices values changed?

Who has it already, who needs it?

Have you tried it for non-linear system, which?

While Kalman filter is quite very important and useful in many cases, by contradiction let's prove/disprove, What is the list of all similar or better states estimation(observance) methods/blocks?

Have you tried NARX system estimation?

If you fully have the system dynamics governing equations you may no need at all Kalman Filter in your controller implementation (realization) on board hw in C (controller will push control signals to real world plant taking it's output and states from internally implemented pre-estimated your real world plant model), right?

For what do you need specifically EKF and not UKF?