Monitoring and Stabilization of the Fully Automatic Robotic Sensor Assembly Line in the Conditions of Digital Twins

Abstract

The article presents the design and expected results from the use of a digital twin monitoring and management system. It is a continuation of the multi-step improvement of the sensor assembly line until its full automation. Statistical control, machine learning, deep machine learning, fuzzy logic and neural networks have been applied to improve and optimize data selection and analysis. A model of digital twins is applied, in which the analysis and feedback for automatic optimal control are outside the work center of a server or cloud. With this powerful management model, many work centers are serviced. Thus, with the development of a single powerful analysis software from the machine manufacturer, many work centers can be managed. This makes the process more automatic and with a lower cost. A specific goal has been set to speed up the work of the automatic production center for assembling sensors and achieve a production cycle of 5 seconds with improved quality results. This is achieved by optimizing automatic visual control using digital twin analysis. Quality improvement has been achieved by improving the inputs of the assembly process using digital twin analysis.