Robust performance of uncertain systems by output feedback (Q1363792)

The authors formulate and solve a robust linear-quadratic control problem for dynamic systems in which the uncertainties are norm-bounded and appear in all the state, input and output matrices. The uncertainties do not satisfy the general matching conditions and the states are not all accessible for measurements. An output-feedback law is constructed using a parametric algebraic Riccati equation to guarantee the quadratic stability. A linear matrix inequality is formulated. From it, the gain of the robust controller can be derived. An efficient computational algorithm for the linear matrix inequality is developed. The developed theory is illustrated on a double-effect evaporator model.