User talk:Mtxu
From Wikimization
(New page: Hello, I have a question in Control theory: I'm Maria Calle, I have an question abot CVX and how to work with. I hope here is the place for answer about. I'm one student from Spain. I...) |
|||
Line 1: | Line 1: | ||
Hello, | Hello, | ||
- | I | + | I'm Maria Calle, |
- | + | I have a question about [http://www.stanford.edu/~boyd/cvx CVX] and how to work with. | |
- | I'm | + | I'm a student from Spain. I'm studying engineering in automatic control. I'm in my last year and I'm making a project of investigacion. |
+ | It's about LQR and how to synthesize the weighting matrix Q and R by pole assignment. | ||
+ | |||
+ | To make a program about it, I'm reading the article: Linear Quadratic Optimal Output Feedback Control For Systems With Poles In A Specified Region, Lisong Yuan, Luke E. K. Achenie, Weisun Jiang. [http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.42.68] | ||
+ | |||
+ | To initializate the gain matrix they say that they apply the method described in the article: | ||
- | + | [http://convexoptimization.com/TOOLS/Arzelier.pdf Arzelier ''et al''., Pole Assignment of Linear Uncertain Systems in a Sector Via a Lyapunov-Type Approach] | |
+ | |||
+ | they said that it's necessary to solve by convex programm. | ||
+ | Until last week I didn't knew anything about convex program. | ||
- | + | I have a system, A, B and C. It have to move poles to one region, for it, I make a transform to the original system. (A_alfa, B_alfa, C_alfa). | |
- | ---- | + | I have to know the gain matrix, with is: |
+ | |||
+ | <math>K=R*W^{(-1)}</math> | ||
+ | |||
+ | It's knew that: | ||
+ | |||
+ | <math>v(x)=x'*P*x=x'*W^{(-1)}*x</math>. | ||
+ | |||
+ | <math>x'*(A'*P+P*A)*x-2*x'P*B*K*x<0</math>. | ||
+ | |||
+ | <math>u(t)=-K*x(t)</math>. | ||
+ | |||
+ | |||
+ | <math>Hij(W,R)=W*A_{/alfa i} '+A{/alfa i}*W-B_{/alfa i}*R-R'*B_{/alfa j} '<0</math> | ||
+ | |||
+ | and | ||
+ | |||
+ | <math>W=eye(2,2)*w, w \in{} R^{(2x2)}</math> | ||
+ | |||
+ | <math>R=eye(2,2)*w, w \in{} R^{(4x2)}</math>. | ||
+ | |||
+ | The ecuation must: <math>min f(W,R)=min(p1)+min(p2)</math> | ||
+ | |||
+ | subject to: | ||
+ | |||
+ | <math>w_{nn} \leq{} p1</math> | ||
+ | |||
+ | <math>-p2 \leq{} r_{qn}\leq{} p2</math> | ||
+ | |||
+ | with | ||
+ | |||
+ | <math>n=1,....4, q=1...2.</math> | ||
+ | |||
+ | <math>W=w*eye(4)</math> | ||
+ | |||
+ | <math>R=r*eye(2,4)</math> | ||
+ | |||
+ | and: | ||
+ | |||
+ | <math>W \geq{} \epsilon1*eye(4)</math> | ||
+ | |||
+ | <math>Hij(W,R) \leq{} -\epsilon2*eye(4)</math> | ||
+ | |||
+ | Now the article says that aplaing the convex programm, it posible to solve the ecuation. | ||
+ | |||
+ | They present the next algoritm: | ||
+ | 1) Initialization | ||
+ | |||
+ | <math>l=0, W1=W1=eye(4), R1=R2=zeros(2,4)</math> | ||
+ | |||
+ | 2)Calculate | ||
+ | |||
+ | <math>\lambda(W)=\lambda min(W)</math> | ||
+ | |||
+ | <math>\lambda_{Hij}=\lambda max(Hij(Wi,Ri))</math> | ||
+ | |||
+ | 3)if | ||
+ | |||
+ | <math>\lambda(W) \geq \epsilon1</math> | ||
+ | |||
+ | <math>lambda_H(ij) \leq -\epsilon2</math> | ||
+ | |||
+ | STOP | ||
+ | |||
+ | Else | ||
+ | |||
+ | <math>l=l+1</math> | ||
+ | |||
+ | calculate eigenvalues(v_w y v_Hij). | ||
+ | |||
+ | calculate the constraint linear: C1(W,R) | ||
+ | |||
+ | <math>v_w'*W1*v_w \geq \epsilon1</math> if <math>(\epsilon1-\lambda_w)>(\lambda_Hij+\epsilon2)</math> | ||
+ | |||
+ | <math>v_Hij'*Hij*v_Hij \leq -\epsilon2</math> if <math>(\epsilon1-\lambda_w)\leq (\lambda_Hij+\epsilon2)</math> | ||
+ | |||
+ | 4)Solve | ||
+ | |||
+ | <math>min(p1+p2)</math> | ||
+ | |||
+ | under | ||
+ | |||
+ | <math>\epsilon1 \leq w_{nn} \leq p1</math> | ||
+ | |||
+ | <math>-p2 \leq r_{qn}\leq p2</math> | ||
+ | |||
+ | <math>C_k(W,R)</math> | ||
+ | |||
+ | <math>k=1,...z</math> | ||
+ | |||
+ | Now I have no idea of how to continue, or how to program the algoritm. | ||
+ | |||
+ | They said the method for solving it's mathematical convex program. | ||
+ | |||
+ | Thanks a lot for all, Greetings. |
Current revision
Hello,
I'm Maria Calle,
I have a question about CVX and how to work with.
I'm a student from Spain. I'm studying engineering in automatic control. I'm in my last year and I'm making a project of investigacion. It's about LQR and how to synthesize the weighting matrix Q and R by pole assignment.
To make a program about it, I'm reading the article: Linear Quadratic Optimal Output Feedback Control For Systems With Poles In A Specified Region, Lisong Yuan, Luke E. K. Achenie, Weisun Jiang. [1]
To initializate the gain matrix they say that they apply the method described in the article:
they said that it's necessary to solve by convex programm. Until last week I didn't knew anything about convex program.
I have a system, A, B and C. It have to move poles to one region, for it, I make a transform to the original system. (A_alfa, B_alfa, C_alfa).
I have to know the gain matrix, with is:
It's knew that:
.
.
.
and
.
The ecuation must:
subject to:
with
and:
Now the article says that aplaing the convex programm, it posible to solve the ecuation.
They present the next algoritm: 1) Initialization
2)Calculate
3)if
STOP
Else
calculate eigenvalues(v_w y v_Hij).
calculate the constraint linear: C1(W,R)
if
if
4)Solve
under
Now I have no idea of how to continue, or how to program the algoritm.
They said the method for solving it's mathematical convex program.
Thanks a lot for all, Greetings.