On an input recovery problem in a linear delay system

We consider the problem of recovering the input of a linear differential equation with delay and propose a solution algorithm that is stable to perturbations. The algorithm is based on the extremal shift principle known in the theory of guaranteed control.