Identification of MIMO (Multi-Input Multi-Outout) Nonlinear Systems Using a Forward-Regression Orthogonal Estimator

摘要

An orthogonal least squares estimator, originally derived for single-input single-output systems, is extended to multi-input multi-output nonlinear systems. The estimator can provide information about the structure, or which terms to include in the model, and final parameter estimates in a very simple and efficient manner. Multivariable nonlinear model validity tests are discussed, and the results of applying the orthogonal estimator to both simulated and real data are included. It is shown that this estimator efficiently combines structure determination with parameter estimation and, when coupled with model validity tests, is particularly powerful in identifying parsimonious models for structure-unknown systems.