Performance assessment of non-Gaussian MIMO system based on entropy and mutual information¹

Shichen Jia,Jinglin Zhou

DOI: https://doi.org/10.1109/CAC57257.2022.10055672

2022-01-01

Abstract:In real production, the performance assessment of control loops has an important role. In this paper, the performance evaluation of MIMO systems is transformed into the performance evaluation of multiple single loops. To address the shortcomings of the canonical correlation analysis (CCA) method, the authors propose a principal component extraction algorithm based on entropy and mutual information projection to latent structure (E-PLS). This algorithm allows us to get the latent structure of the coupling MIMO system. Then the latent structure is modeled to construct the virtual system. On this basis, a non-Gaussian minimum entropy approach was used to evaluate the performance of the virtual system.