Identification of piecewise affine systems and nonlinear systems using multiple models

Abstract

In this paper, a procedure for the identification of piecewise affine ARX systems is proposed. The parameters of the individual subsystems are identified through a least-squares based identification method using multiple models. The partition of the regressor space is then determined using standard procedures such as neural network classifier or support vector machine classifier. The same procedure can be applied to identify nonlinear systems by approximating them via piecewise affine systems. Extensive simulation studies show that our algorithm can indeed provide accurate estimates of the plant parameters even in noisy cases, and even when the model orders are overestimated.