Persistent Excitation by Deterministic Signals for Subspace Parametric Identification of MISO Hammerstein Systems

The persistent excitation (PE) issue is addressed for subspace identification of MISO Hammerstein systems. A signal generator is designed that provides persistently exciting inputs for the considered class of systems. Signal generation involves amplitude weighting and phase shifting on a given finite-length discrete-time Dirac Comb signal. It turns out that the generated signals are deterministic and meet the PE requirement within a relatively short time identification horizon. This theoretical feature is confirmed by simulation.