The Frequency-Domain Robust Controller Design Toolbox is a tool for designing robust linearly parameterized controllers in the Nyquist diagram. The method is based on loopshaping with stability and performance constraints and uses essentially linear and quadratic programming (some cases needs Semi-Definite Programming). For linear and quadratic optimization the well-known linprog or quadprog (depending on the problem) commands of the Optimization Toolbox of MATLAB are used. While convex optimization problems are formulated with YALMIP and can be solved with all available solvers. Many commands of the Control Toolbox of MATLAB are used as well.
This toolbox has the following features:
- Full compatibility with model structures of Matlab (LTI with delay, frequency-domain models, discrete- or continuous-time models).
- For Idmodels (from identification toolbox), frequency-domain uncertainty or parametric uncertainty are automatically taken into account for robust controller design.
- Multimodel uncertainty is supported.
- Gain-scheduled controller can be designed.
- H infinity performance can be defined independently for all closed-loop sensitivity functions.
- Fixed-structure controller design (like PID controller) is supported.
- Multivariable decoupling controller can be designed.
In the following version, some bugs are fixed and a GUI is included (May 2014):