IDENTIFICATION OF NONLINEAR-SYSTEM STRUCTURE AND PARAMETERS USING REGIME DECOMPOSITION

An off-line algorithm for empirical modeling and identification of non -linear dynamic systems is presented. The minimal input to the algorit hm is a sequence of empirical data and the model order. Using this inf ormation, the algorithm searches for an optimal model structure and pa rameters within a rich non-linear model set. The model representation is based on the interpolation of a number of simple local models, wher e each local model has a limited range of validity, but the local mode ls yield a complete global model when interpolated. The method is illu strated using simulated data.