Add discrete-time steady-state Kalman filter

[wei-chen-li]

This PR implements the discrete-time steady-state Kalman filter and Luenberger observer.

For reference, the Kalman filter dynamics is given by:

$$ K = \hat{P} C^\top (C \hat{P} C^\top+ V)^{-1} $$ $$ \hat{x}[n|n] = \hat{x}[n|n-1] + K(y - \hat{y}) $$ $$ \hat{x}[n+1|n] = A \hat{x}[n|n] + B u[n] $$

At steady-state, the dynamics is given by:

$$ \hat{x}[n+1|n] = A \hat{x}[n|n-1] + B u[n] + L(y - \hat{y}) $$

Thus, the observer gain is:

$$ L = A \hat{P} C^\top (C \hat{P} C^\top + V)^{-1} $$

---

This change is 

[jwnimmer-tri]

+@RussTedrake for feature review or delegation, please.