Nonlinear disturbance decoupling and linearization: a partial interpretation of integral feedback

The relation between the solvability of the disturbance decoupling problem for a nonlinear system and its linearization around a working point is investigated. It turns out that generically the solvability of the disturbance decoupling via regular dynamic state feedback is preserved under linearization. This result gives a partial interpretation of introducing integral action in classical PID-control applied to nonlinear systems. The theory is illustrated by means of a worked example