Stochastic subspace algorithm based on the orthogonal decomposition method for closed-loop system identification

This paper presents a stochastic subspace identification algorithm dedicated to subspace-based closed-loop system identification. The preliminary algorithm is based on the ordinary subspace technique followed by a modification in which a new proposed method based on orthogonal decomposition method is used to reconstruct the past input and past output data of the instrumental variables. The noise model is obtained as to improve the performance of the model in dealing with a stochastic system. The analysis of the proposed approach with PO-MOESP method is also carried out. The efficacy of the developed approach is then demonstrated by identifying a simulation based on experimental data. An evaluation performance test according to mean square errors, variance accounted for and best fit, are used as to verify the accuracy of the model. A comparative simulation results with PO-MOESP method is also included.