Control Bootcamp: Observability

This video explores the observability of a linear system, namely the ability to estimate the full state "x(t)" from a time-history of limited output measurements "y(t)".

These lectures follow Chapters 1 & 3 from:

Machine learning control, by Duriez, Brunton, & Noack

This video was produced at the University of Washington

[Drone.Robotics]

Thank you Sir....I have seen the whole series and it have cleared lot of my concepts about control theory. Your videos are just great and your way of teaching complex things in simple manner is appreciable. Thanks Again.

[jakesnakebakecake]

Hey professor, at 2:00 shouldn't the last column of the controllability matrix be A^(n-1)*B?

[anhtieng]

what if we have a system that the controllability matrix rank is less than n and observability matrix rank is n... That means, we would be able to estimate full state x but cant do anything to control the system... Does this case ever happen pls?

[zrmsraggot]

can observability be used for system identification ? Having input and output but not knowing about the dynamics

[AchrajSarma]

When you say system is "more observable/controllable in some direction", how does that translate to which state is better estimated/controlled? With SVD, I will find the vector specifying that direction, but how do I connect it to the corresponding element in the state vector (is it X1, or X2, or so on...)?

[Jeff-zcrr]

Another definition i read is "a system is observable if you can calculate it's initial state from it's measurements." This definition sounds a bit more intuitive.

[Jeff-zcrr]

Is it possible to explain Observability without using th word observable?