System Identification using Delaunay Tessellation of Self-Organizing Maps

The Self-Organizing Map (SOM) is a vector quantization method which places prototype vectors on a regular low-dimensional grid in an ordered fashion. A new method to obtain piecewise linear models of dynamic processes is presented. The operating regimes of the local linear models are obtained by the Delaunay tessellation of the codebook of the SOM. The proposed technique is demonstrated by means of the identification of a pH process.