Comparison of subspace-based system identification techniques

In this paper, we applied several types of predictor-based system identification technique to various systems such as electronics, process industry, biomedical, and heating system in order to compare its performance. The PBSID technique shows the best performance with consistent identification, which is based on the Kalman observer calculated from the Riccati equation. The observer matrix is guaranteed to be stable even the original system is unstable regardless with the fact that the identified system is under open or closed loop. Its effectiveness is demonstrated with simulation using the Matlab^©.