Decoupling Nonlinear Models of the F-16 Aircraft Benchmark

System identification is witnessing a paradigm shift from linear to nonlinear modeling. Several nonlinear models have been studied, ranging from white-box models that take represent the governing physics, over gray-box model structures where a specific parameterization is considered, to black-box models which contain numerous parameters or may even be nonparametric. The model classes in the darker shades of gray have more descriptive power than the lighter and structured ones, but they come at a price. They typically have a high parametric complexity, and are more difficult to interpret. For instance, when using polynomials as basis functions, the number of terms of increases combinatorially with the number of variables i.e., regressors in a NARX model, or states in a statespace model), as with the nonlinear degree.