Fine Tuning Kalman Filter Using Simulink Design Optimization

I would like to tune my Kalman filter (Q and R matrices) using design optimization toolbox. I introduce the estimated signal and add a signal property to the optimizer. The signal property is Track Reference Signal with proper Time Vector and Amplitude. Although, the initial values of Q and R are relatively good, the second estimated signal ends up being zero. I wonder how I can prevent that, or in general tune my filter more optimaly.

Answer 1

[Kaashyap Pappu]

A similar query regarding the Q and R matrices’ optimization was addressed and elaborated upon here. It is important to model the noise characteristics prior to Kalman Filtering to obtain the Q and R matrices for better performance.