On the plant augmentation by integrators in the discrete-time LQG/LTR control

This paper presents a study on the peculiarities of the plant augmentation by integrators in the discrete-time LQG/LTR control. In the LQG/LTR methodology, the inclusion of “free” integrators in each control channel of the plant helps the designer to define a Target Feedback Loop with good performance characteristics in the mixed-sensitivity analysis. However, due to specific conditions of the discrete-time case, the integration method used in the plant augmentation can make the application of the LTR principle unfeasible. In this context, an analysis of the feasibility of the augmentation of the plant by forward Euler and backward Euler integrators is presented. Also, a new Target Feedback Loop parametrization is presented that allows the bound on the sensitivity function of the control loop to be associated with the behavior of the inverse of an integrator, ensuring good properties of disturbance rejection and tracking of reference signals.