Invertibility of sampled data systems

In this paper, we study the invertibility of a sampled data system. Necessa ry and sufficient invertibility conditions are given explicitly in terms of transfer function zeros. It is shown that many systems including those wit h "unstable" zeros or those with more inputs than outputs are in fact stabl y invertible. Detailed design procedures to achieve the system inverse are given. A minimum sensitivity Inverse with respect to under-modeling and noi se is also provided in the paper.