The MAS4AI framework for human-centered agile and smart manufacturing

Area of Science:

• Industrial Engineering
• Artificial Intelligence
• Cyber-Physical Systems

Background:

• Modern value chains require adaptable, resilient production systems to meet demands for customized, low-batch products.
• Cyber-Physical Production Systems (CPPS) with Artificial Intelligence (AI) can enhance system resilience and responsiveness.
• Existing multi-agent solutions face challenges in orchestrating AI components and integrating human decision-making.

Purpose of the Study:

• To propose a Multi-Agent System (MAS) for dynamic and reconfigurable production paradigms.
• To enhance human-machine integration and collaboration within intelligent manufacturing systems.
• To enable autonomous and cooperative decision-making through decentralized agent interactions.

Main Methods:

• Development of a Multi-Agent System (MAS) where software agents collaborate with smart workers.
• Utilization of Asset Administration Shell (AAS) submodels for machine-readable smart worker descriptions.
• Implementation of AAS for system integration, adaptability, and self-configuration.

Main Results:

• The proposed MAS facilitates dynamic and reconfigurable production environments.
• Asset Administration Shell (AAS) enables seamless integration and self-description of system components.
• Improved human-machine integration in shared assembly tasks and support for plug-and-produce functionality.

Conclusions:

• The MAS approach enhances production system resilience and adaptability.
• AAS integration is crucial for interoperability and dynamic reconfiguration.
• The system promotes efficient human-machine collaboration in Industry 4.0 settings.