Controller performance assessment for switched linear systems via applying non-negative tensor factorization

This paper considers the controller performance assessment problem based on the data-driven benchmark by applying the tensor approach for switched systems. For each subprocess, the multivariate output with finite sample numbers is extracted regarding as the forms of feature matrices. Consequently, by stacking the feature matrices associated to any data set samplings, a tensor is created, a fact which necessitates studying the controller performance assessment for switched systems considered using tensors. In order to fit the multiple invariance of the measurement output tensor data processing for controller performance assessment, the non-negative tensor factorization model is proposed to analyze the tensor data, which stems from uniqueness of low-rank decomposition of higher-order tensor. By using the non-negative tensor factorization, the estimated latent outputs structure combining with the covariance-based algorithm are given to derive the data-driven performance benchmark and performance index for each subprocess controller, and the presented data-driven benchmark estimation algorithm requires a set of close-loop routine operating data. A simulation example is provided to test the effectiveness and advantages of proposed method.