Domain modeling for scenario sensing and edge decision-making

Abstract

The introduction of numerous edge computing nodes allows application systems to sense and make decisions in real-time but also brings new challenges. The diversity of application scenarios and the complexity of application processes can be effectively addressed through modeling. This paper proposes a modeling approach for manufacturing scenario sensing and edge decision-making. Firstly, an abstract meta-model (SMM) is defined, which provides a unified description of the resources and processes in the scenario and the interaction between the scenario and the edge. Based on the meta-model, an application scenario model (ASM) can be constructed for a specific scenario to support edge data feedback and decision-making for abnormal events. In addition, the model is constructed in a scenario modeling tool and validated in a simulated manufacturing production line, that is, whether the models can provide effective support for decision-making of abnormal events. The results demonstrate that mapping normalized models into codes at the edge computing nodes can improve the accuracy and real-time performance of decision-making.