Closed-loop identification revisited (Q1301435): Difference between revisions

The paper deals with the identification of multiple-input multiple-output (MIMO) systems operating in closed loop under, in general, linear, time-invariant feedback. The leading idea in the paper is to provide a unified framework for the prediction error methods. The paper locates the common approaches in a coherent setting and shows the connections, similarities and differences in the asymptotic behaviour of the particular estimates. It is shown that most of the methods correspond to different parameterizations of the dynamics and noise models. In effect, specific asymptotic statistical properties of the estimates are derived from the general convergence theory for particular parameterizations. A comparative study of the different methods is given and a summarizing discussion is presented.