TV-LQG notebooks

[gchenfc]

A series of notebooks to walk through time-varying LQG.

1. LQR
2. Kalman Filter gains + LQR = LQG
3. Extending to Nonlinear: TV-LQG

[dellaert]

Hey @gchenfc , you do know about the existing LQR notebooks, right?

• https://borglab.github.io/gtsam/lqrexample/
• https://borglab.github.io/gtsam/ilqrexample/
  They already built on your blog post, comments/edits very welcome!
  So please refer to those notebooks and continue the series with that background information perhaps?
  I'm also asking Andrew, your successor ;-), to take a look at your notebook and try to run it. I know this is a draft PR but maybe early feedback would be welcome.

[gchenfc]

Oh I did not know about those! Biggest comment I think is that I think both the LQR notebook and iLQR notebook should show how to extract the gain matrices from the bayes net, but maybe I just missed that (I'm on mobile atm).

My notebook so far only has the LQR part (I only got through "part 1" so far), so will probably get axed according to your suggestion to start out referencing the iLQR & LQR notebooks and proceeding from there.

[TingWeiWong]

Hi Gerry, awesome blog post (https://gtsam.org/2019/11/07/lqr-control.html) on LQR! There’s a small typo in the graph. $$Ax_1 + Bu_1 = X_x$$ should be $$Ax_1 + Bu_1 = X_2$$ instead.

[TingWeiWong]

Rendering error for bold $\bm x, \bm u$?