Instrumental variables for nonlinearity recovering in block-oriented systems driven by correlated signals

摘要: The goal of the paper is to identify Hammerstein-type systems excited and disturbed by correlated random processes. problem semi-parametric in sense that nonlinear static characteristic recovered without prior knowledge about linear dynamic block, i.e. when its order unknown. method based on instrumental variables technique, with instruments generated input inverse filtering. It proved that, contrast least-squares-based approach, proposed algorithm leads an asymptotically unbiased, strongly consistent estimate. Constructive procedures generation are given for some popular cases.