Тёмный

MAE509 (LMIs in Control): Lecture 9 - H-infinity optimal Full-State Feedback 

Cybernetic Systems and Controls
Подписаться 3,8 тыс.
Просмотров 12 тыс.
50% 1

In this short lecture, we combine the LFT, the KYP Lemma, Schur complement, Duality, and Variable Substitution to find an LMI for Hinfinity optimal full-state feedback controller synthesis.

Опубликовано:

 

11 сен 2024

Поделиться:

Ссылка:

Скачать:

Готовим ссылку...

Добавить в:

Мой плейлист
Посмотреть позже
Комментарии : 19   
@YLlh9
@YLlh9 9 месяцев назад
You are execllently good Professor!
@faridoudjama3104
@faridoudjama3104 3 года назад
Sir. Can you give me code matlab H-infinity optimal Full-State Feedback usingLinear Matrix Inequality?
@tanaykumar3493
@tanaykumar3493 Год назад
Sir the 2nd point in Theorem 5 is an LMI in two variables which are Y and Z. I wanted to know if interpreted that correctly or not? And if that is correct how do we solve for two variables with just one equation?
@MatthewPeet
@MatthewPeet Год назад
Correct! Any LMI solver will allow you to declare multiple variables and multiple constraints (there are two here -- Y>=0 and []
@tanaykumar3493
@tanaykumar3493 Год назад
@@MatthewPeet thank you sir for your help!
@hectorgautier4061
@hectorgautier4061 9 месяцев назад
Mister, thanks a lot for such riveting content. If I may ask a question, how can we know the shape and entries' value of B1 and D11? I do not really get the meaning of D11. Is it free decisionning for the conceptor? Based on its position in the KYP matrix, D11 only constraints the objective variable gamma's value and not Y and Z, is not it? About B1, I suppose that its shape depends on the type of disturbances we are considering. However, what about its entries' value? Thanks a lot.
@MatthewPeet
@MatthewPeet 9 месяцев назад
D11 is the direct feed-through term. It is significant if the disturbance is measured directly in the regulated output, as is the case in the tracking framework, where e(t)=y(t)-r(t) is regulated and r(t) is the reference signal which we treat as a disturbance. The number of rows in B1 is the number of states, and the number of columns is the number of disturbances. The values will depend on the system you are modelling, so its hard to say a priori.
@afaq5021
@afaq5021 Год назад
If we have control constraints like "u" should be between some range. How can we approach that?
@MatthewPeet
@MatthewPeet Год назад
Because we are looking for K here and not u, it doesn't fit directly in the controller synthesis framework. Two typical options: 1) formulate directly as an optimization problem in terms of (Lecture 2), e.g. dynamic programming or MPC. 2) weight your actuation signal heavily in the regulated output and hope for the best.
@richitrx7
@richitrx7 2 года назад
Hello Dr. Peet, is it possible to derive an LMI where both u and w use static state-feedback (u=Fx, w=Lx), I've seen hamiltonian derivations for the solution. But have not been successful in forming such an LMI, pure curiosity.
@richitrx7
@richitrx7 2 года назад
without the requirement of the system matrix (A) being Hurwitz
@MatthewPeet
@MatthewPeet 2 года назад
w is a disturbance, so that wouldn't make sense. Maybe you mean something about estimator design?
@richitrx7
@richitrx7 2 года назад
@@MatthewPeet I think it might be used to simulate the worst disturbance, as reference "Robust and Optimal Control" by Zhou, Doyle, & Glover pg 463
@mohammedyaqoob9154
@mohammedyaqoob9154 2 года назад
Hello dear doctor I have a state space model with 2 equations in this form (X(k+1)=Ax+dw+Bu and Y(k)=Cx+Du), so, please,how can represente the 9 matrices form that?
@MatthewPeet
@MatthewPeet 2 года назад
Well, you need a regulated output. Otherwise its a 6-matrix representation.
@mohammedyaqoob9154
@mohammedyaqoob9154 2 года назад
@@MatthewPeet thank you for your reply Dr🌹🙏🌹
@mohammedyaqoob9154
@mohammedyaqoob9154 2 года назад
@@MatthewPeet but please, Dr. Do you have a matlab code for example that design H infinity controller with uncertainty in lmi because I don't deal with robust Control before??
@MatthewPeet
@MatthewPeet 2 года назад
@@mohammedyaqoob9154 check out the course wikibook!
@mohammedyaqoob9154
@mohammedyaqoob9154 2 года назад
@@MatthewPeet thank you so much Dr. 🙏
Далее
H Infinity and Mu Synthesis | Robust Control, Part 5
13:57
Control Bootcamp:  Kalman Filter Example in Matlab
22:12
L34B:  The State Feedback H∞ Control
9:27
Просмотров 18 тыс.
Disk Margins for MIMO Systems | Robust Control, Part 3
15:57
Control Design via State space
38:03
Просмотров 77 тыс.
Control Bootcamp:  Introduction to Robust Control
8:13
Model Predictive Control
12:13
Просмотров 256 тыс.