How can I use HINFSTRUCT in order to find minimal-norm static-output-feedback?
Specifically, if A, B and C are the state space matrices of a continuous time linear system, how can I find, using HINFSTRUCT, a matrix K, such that A-BKC would be stable and have a minimal norm?
Also, if the system is a discrete time system, can I use HINFSTRUCT in the same way?
According to my experience Matlab (Robust Control Toolbox) is the best way to design such a controller.
You are able to find the subroutine on the Help files easily.
Best
I've written some comments and attached here. Please refer to the attached file.
As mentionned by Mohammad Gudarzi, the help in Matlab is pretty well documented. You also can find some examples on Pierre Apkarian pages (see link).
As far as the code is concerned, you can use specific variable types like realp to create tunable gains. Expected requirements are expressed in the form of weightings on a user-defined synthesis model. Note that a static gain may not be sufficient to enforce these expected requirements depending on the model. You can consider other structures (like observer-based in the case of output feedback) which can in turn be tuned with HINFSTRUCT.
The optimization algorithm returns the minimal achieved H-inf norm which helps decide wether requirements are too strong or the number of degrees of freedom is not big enough.
Best regards.
http://pierre.apkarian.free.fr/EXAMPLES_STRUCTURE/ContentHTML/Content.html
http://www.mathworks.fr/help/robust/ref/hinfstruct.html;jsessionid=eafae802f7c501aaa4c487c790bc
- Badges
- Science topic