Identification of errors-in-variables ARX models using modified dynamic iterative PCA

Identification of autoregressive models with exogenous input (ARX) is a classical problem in system identification. This article considers the errors-in-variables (EIV) ARX model identification problem, where measurements are also corrupted noise. The recently proposed Dynamic Iterative Principal Components Analysis (DIPCA) technique solves EIV but only applicable to white measurement errors. We propose novel algorithm based on modified DIPCA approach for identifying EIV-ARX single-input, single-output (SISO) systems output coloured noise consistent model. Most existing methods assume important parameters like input-output orders, delay, or noise-variances be known. work’s novelty lies joint estimation error variances, process order, and parameters. central idea used obtain all these theoretically rigorous manner transforming lagged using appropriate covariance matrix , which obtained estimated variances Simulation studies two presented demonstrate efficacy algorithm.