System Identification: Theory for the User, 2nd Edition

5. Models for Time-varying and Nonlinear Systems

While linear, time-invariant models no doubt form the most common way of describing a dynamical system, it is also quite often useful or necessary to employ other descriptions. In this chapter we shall first, in Section 5.1, discuss linear, time-varying models. In Section 5.2 we deal with models with nonlinearities in form of static, non-linear elements at the input and/or the output. We also describe how to handle nonlinearities that can be introduced by suitable nonlinear transformations of the raw measured data. In Section 5.3 we describe nonlinear models in state-space form. So far the development concerns model structures, where the nonlinearities are brought in based on some physical insight. ...