Output regulation problem

摘要

This paper deals with the design of an optimal error feedback regulator for linear t-invariant stochastic controlled system. The presented solution of output regulation problem is based on the "observed-based" approach and removes known restrictions of the problem formulation. Kalman canonical decomposition of controlled system is used for objective definition of an augmented plant containing plant and exo-system. The deviation model of the augmented plant simplifies analysis and design of an error feedback regulator. Then standard tools of LQG optimization and Kalman filter design are used. It is shown that the obtained error feedback regulator can be interpreted as multivariable PFD regulator. This regulator contains derivation, proportional and compensation blocks equivalent to the exosystem under the internal model principle.