A Deep Neural Network with Module Architecture for Model Reduction and its Application to Nonlinear System Identification

Abstract

This paper proposes a deep neural network with module architecture for model reduction, and a cost function suitable for training the model. In the proposed model architecture, each layer is modularized to reduce the model by adjusting the number of layers. This feature allows the computational load of the model to be quickly adjusted. In order to maintain the accuracy of the reduced model even if it is not retrained, the cost function is defined as a weighted average of the errors of the model output over the number of modules. Finally, this architecture is incorporated into nonlinear Linear Fractional Representation (LFR) models for nonlinear system identification. The effectiveness of the proposed method is illustrated through numerical examples.