Linsys lab dag 4

Changing $Q_d$ (which is the covariance of the stochastic disturbance):

• Higher values of $Q_d$ means that its high uncertainty in the model dynamics, and will trust the measurements more.
• Lower values of $Q_d$ means that the model is very accurate, and we will trust the measurements less.

We have done following measurements in the lab:

1. Way too high $Q_d$ matrix (10 on all diagonal elements)
2. Way too low $Q_d$ matrix (0.000001 on all diagonal elements)
3. Decently good $Q_d$ matrix (0.0001 on all diagonal elements) 3.1 Losing new data (with an implemented switch), and seeing how the system reacts to this.

We observed this

1. .
2. .
3. . 3.1