Data-Driven Control: Observer Kalman Filter Identification
Ғылым және технология
In this lecture, we introduce the observer Kalman filter identification (OKID) algorithm. OKID takes natural input--output data from a system and estimates the impulse response, for later use with the eigensystem realization algorithm (ERA).
www.eigensteve.com/
This video was produced at the University of Washington
Пікірлер: 8
I don't know how grateful you are to provide this course for free. I'm from Korea, and studying control engineering by my self. Thanks a lot Professor.
Amazing lecture ⬆️⬆️⬆️⬆️
Hi Brunton, Your lectures made it much simpler a process for me to understand these complex algorithms with little ease. Can you plz help me in understanding direct data driven online control schemes. Like model free control schemes... I really want to learn them through your videos. thank you.
Great video, loved it! Just a question: Is it possible to derive a Kalman filter for state estimation by using the Markov parameters identified by OKID?
I am of the understanding that we can find an impulse response for a linear system from input-output data by solving for x in Ax=y, where A is a matrix of past samples of the input and y is a vector of outputs (i.e linear regression). What is the advantage of OKID over this approach?
@gymzatan7824
Жыл бұрын
I think it would have to do with the bias and variance. For example, put a normally distributed variable in the denominator and you’ll get a skewed outcome. Same with matrix inversion. So you’ll need some way to adjust to that bias.