Symbolic and numerical computations for analysis of uncertain systems

Analysis of control systems with respect to parametric uncertainties is the subject of this paper. Various closed loop and MIMO structures containing uncertain parameters are considered. We show how symbolic and numerical computations can cooperate to do this job, i.e., the evaluation of the particular system function is performed symbolically. This is necessary since many unknown entries are present. However, analysis of the result can be done numerically and efficiently. Related software tools we call for include: the Symbolic Math Toolbox for Matlab, and the Polynomial Toolbox for Matlab with its powerful tools for studying parametric systems. This combination appears fruitful and has been applied successfully to robust control techniques at the Czech Technical University in Prague.