ADAPTABILITY IN INDUSTRY 4.0: SERVICE-ORIENTED ARCHITECTURE TO DEPLOY ARTIFICIAL INTELLIGENCE ON INDUSTRIAL AUTOMATION

Abstract

Industry 4.0 represents a revolution in the business environment, driving the integration of information technologies and industrial automation, with the main objective of reducing latency in decision-making. Artificial Intelligence (AI) plays an essential role in advanced data analysis and resource optimization, allowing accurate predictions and agile decisions, however its implementation presents challenges, such as algorithm complexity and integration with industrial automation systems. An innovative solution to overcome these challenges is the implementation of a service-oriented architecture, which creates modular and interoperable systems, a concept especially relevant in the practical application of industrial automation, in which the integration between Information Technology (I.T) and Automation Technology (A.T) is crucial. This work presents, through experimental research, an innovative solution based on the development of a computer vision application, isolated in a Docker container. This application is designed to inspect the assembly of parts by a robotic system and establish communication with a Programmable Logic Controller (PLC) to approve or disapprove the assembly. The results of the adopted architecture demonstrate a flexible approach that simplifies the operation of AI systems, allowing operation with both AI enabled and disabled, reducing potential disruptions to the workflow. This research promises to open paths for future innovations and advances in the field of industrial automation and Artificial Intelligence, offering a model that effectively combines the agility of AI with the robustness of automation.