Agent-Based Approach for Decentralized Data Analysis in Industrial Cyber-Physical Systems.

The 4th industrial revolution is marked by the use of Cyber-Physical Systems (CPSs) to achieve higher levels of flexibility and adaptation in production systems that need to cope with a demanding and ever-changing market, driven by mass customization and high quality products. In this context, data analysis is a key technology enabler in the development of intelligent machines and products. However, in addition to Cloud-based data analysis services, the realization of such CPS requires technologies and approaches capable to effectively support distributed and embedded data analysis capabilities. The advances in Edge Computing have promoted the data processing near or at the devices that produce data, which combined with Multi-Agent Systems, allow to develop solutions based on distributed and interacting autonomous entities in open and dynamic environments. In this sense, this paper presents a modular agent-based architecture to design and embed cyber-physical components with data analysis capabilities. The proposed approach defines a set of data processing modules that can be combined to build cyber-physical agents to be deployed at different computational layers. The proposed approach was applied in a smart inspection station for electric motors, where agents embedding data analysis algorithms were distributed among Edge, Fog and Cloud layers. The experimental results illustrated the benefits of distributing the data analysis by different computational layers.