A multiagent based knowledge extraction framework to support plug and produce capabilities in manufacturing monitoring systems

Abstract

The manufacturing industry has been steadily evolving over the years, with new market trends encouraging manufacturers to find new ways to meet the consumers' demands and quickly adapt to new business opportunities. Manufacturing systems are therefore required to be more and more agile and flexible in an environment dominated by unpredictable changes and disturbances. As a direct consequence several new solutions have been proposed, revolving around agility, flexibility, reconfigurability and modularity, enabling concepts such as Plug & Produce (P&P). Following this trend, the present article proposes a possible implementation for a multiagent-based knowledge extraction architecture to support P&P in flexible, distributed manufacturing monitoring systems. The validation process is also described, entailing the application of said system in a real industrial environment, more specifically monitoring two robotic cells performing the welding of a car's side member.