Output feedback stabilization of uncertain linear systems

The problem of static output feedback stabilization is solved for continuous- and discrete-time linear systems subject to possibly time-varying bounded parametric uncertainties. No statistical description of the uncertainties is assumed. Using Lyapunov stability considerations, the controller design problem is formulated as an equivalent smooth unconstrained function minimization problem, which is then solved by standard gradient-based algorithms. Performance specifications concerning a quantity defined as the degree of exponential stability can be naturally considered in the design process. Analytical expressions for the gradients of the function to be minimized are given. An example which allows comparison with other design methods is presented, and several practical aspects are discussed