Using a Wiener-Type Recurrent Neural Network with the Minimum Description Length Principle for Dynamic System Identification

Abstract

This paper presents a novel Wiener-type recurrent neural network with the minimum description length (MDL) principle for unknown dynamic nonlinear system identification. The proposed Wiener-type recurrent network resembles the conventional Wiener model that consists of a dynamic linear subsystem cascaded with a static nonlinear subsystem. The novelties of our approach include: 1) the realization of a conventional Wiener model into a simple connectionist recurrent network whose output can be expressed by a nonlinear transformation of a linear state-space equation; 2) the state-space equation mapped from the network topology can be used to analyze the characteristics of the network using the well-developed theory of linear systems; and 3) the overall network structure can be determined by the MDL principle effectively using only the input-output measurements. Computer simulations and comparisons with some existing recurrent networks have successfully confirmed the effectiveness and superiority of the proposed Wiener-type network with the MDL principle.