Dynamic inversion of linear multivariable systems with observers

This paper describes the structure of the static state-feedback compensator of dynamic inversion for general linear multi-input multi-output systems. For a large class of parameter-dependent systems, dynamic inversion removes the parameter dependency from the input-output map and gives a time invariant decoupled system. The use of an observer does not affect the decoupling by dynamic inversion. The effect of an observer amounts to injecting an uncontrollable observer error signal at the input of the decoupled system.