摘要：AbstractIdentification of closed-loop systems from the input-output data has been studied quite extensively for several decades. In this work, we extend the use of the dynamic iterative principal components analysis (DIPCA algorithm by Maurya et al.) to the task of identification of the model in closed-loop, where the input and output are corrupted with heteroscedastic white-noise also known as the errors-in-variables class of problems. We develop and evaluate the DIPCA approach for two methods, viz., the direct method and the two-step method for closed-loop system identification. Monte Carlo simulation results are presented to demonstrate the consistency of the results.
