43 - Realization of discrete-time periodic systems from input-output data

Abstract

This chapter presents a method for the direct realization of a linear periodically time varying state-space model from input-output data. The algorithm is based on an implicit reformulation into several time-invariant monodromy systems, derived in an operator theoretic framework. The singular value decomposition (SVD)-exploiting subspace algorithm is extended to the periodic case. The chapter focusses on the exact and fortuitous reconstruction of the system from noiseless data and shows that the algorithm performs well in moderately noisy conditions.