Critical Logical Link Assessment for Improving Industrial Wireless Communication Systems

Wireless communication is an important power for moving Industry 4.0 and intelligent manufacturing, where its dependability is challenging industry and academia. Dependability research has two phases: dependability assessment and control. Dependability control based on assessment feedback is the ideal mode for future complex industrial wireless communication systems (WCSs). However, the state-of-the-art automated control methods are focused on improving dependability indicators which ignore the boundary that only killing some logical links may improve the whole WCS. Therefore, this paper proposes the critical logical link assessment method with steps of measurement, experience-based labeling, feature extraction, link assessment, and critical link calculation. In particular, the feature extraction and link assessment are based on two deep neural networks and work together to obtain a correlation coefficient between two logical links. At last, we conducted experiments on realistic measurements in three WCSs which proves both usability and novelty.