Chapter 6. System Identification Methods

6.1 Identification Methods Based on the Whitening of the Prediction Error(Type I)

6.1.1 Recursive Least Squares (RLS)

6.1.2 Extended Least Squares (ELS)

6.1.3 Recursive Maximum Likelihood (RML)

6.1.4 Output Error with Extended Prediction Model (OEEPM)

6.1.5 Generalized Least Squares (GLS)

6.2 Validation of the Models Identified with Type I Methods

6.3 Identification Methods Based on the Uncorrelation of the Observation Vector and the Prediction Error (Type II)

6.3.1 Instrumental variable with Auxiliary Model (IVAM)

6.3.2 Output Error with Fixed Compensator (OEFC)

6.3.3 Output Error with (Adaptive) Filtered Observations (OEAFO)

6.4 Validation of the Models Identified with Type II Methods

6.5 Estimation of the Model Complexity

6.5.1 An Example

6.5.2 The Ideal Case (no Noise)

6.5.3 The Noisy Case

6.5.4 Criterion for Complexity Estimation

6.6 Concluding Remarks

6.7 Notes and References