TY - GEN
T1 - Industrial Artificial Intelligence
T2 - 20th IEEE International Conference on Industrial Informatics, INDIN 2022
AU - Salazar, Luis Alberto Cruz
AU - Vogel-Heuser, Birgit
N1 - Publisher Copyright:
© 2022 IEEE.
PY - 2022
Y1 - 2022
N2 - 'Artificial Intelligence in Industry 4.0', a technical report published by the working groups 'Technological and Application Scenarios' and 'Artificial Intelligence' (AI) of the Industry 4.0 (I4.0) platform, presents an innovative Industrial AI concept. Above all, it concludes that I4.0 experts and scientists must become accustomed to the behavior of autonomous AI-controlled systems, collaborate with them and comply with learnability requirements (predictability). Industrial AI instantly raises a set of concerns about existing norms and new standardizations. These frequently provide guidelines and, in some cases, offer procedures and implementations using design patterns. One way to produce AI in I4.0 systems is through Industrial Agents (IAs) due to their natural autonomy and additional intelligent characteristics, e.g., reactiveness, proactiveness, and human cooperativeness. Multi-Agent Systems (MASs) are particularly well suited for representing distributable AI that can develop I4.0 components being applied to various I4.0 scenarios. Considering the properties of IAs and the corresponding standards, an MAS architecture is used to understand the aspects of the flexible, intelligent, and automated Cyber-Physical Production System (CPPS). This article proposes a predictive IA for I4.0 (Agent4.0) to an agent-based CPPS architecture, leveraging IA design patterns and logical structure for implementing MAS. As a result, relevant standardized IA design patterns for I4.0 show how MAS can be created with the help of the Industrial AI requirements and Agent4.0 skills (functions) identified.
AB - 'Artificial Intelligence in Industry 4.0', a technical report published by the working groups 'Technological and Application Scenarios' and 'Artificial Intelligence' (AI) of the Industry 4.0 (I4.0) platform, presents an innovative Industrial AI concept. Above all, it concludes that I4.0 experts and scientists must become accustomed to the behavior of autonomous AI-controlled systems, collaborate with them and comply with learnability requirements (predictability). Industrial AI instantly raises a set of concerns about existing norms and new standardizations. These frequently provide guidelines and, in some cases, offer procedures and implementations using design patterns. One way to produce AI in I4.0 systems is through Industrial Agents (IAs) due to their natural autonomy and additional intelligent characteristics, e.g., reactiveness, proactiveness, and human cooperativeness. Multi-Agent Systems (MASs) are particularly well suited for representing distributable AI that can develop I4.0 components being applied to various I4.0 scenarios. Considering the properties of IAs and the corresponding standards, an MAS architecture is used to understand the aspects of the flexible, intelligent, and automated Cyber-Physical Production System (CPPS). This article proposes a predictive IA for I4.0 (Agent4.0) to an agent-based CPPS architecture, leveraging IA design patterns and logical structure for implementing MAS. As a result, relevant standardized IA design patterns for I4.0 show how MAS can be created with the help of the Industrial AI requirements and Agent4.0 skills (functions) identified.
KW - Agent4.0
KW - Artificial Intelligence
KW - Cyber-Physical Production Systems
KW - Industrial Agents
KW - Industry 4.0
UR - http://www.scopus.com/inward/record.url?scp=85145776728&partnerID=8YFLogxK
U2 - 10.1109/INDIN51773.2022.9976159
DO - 10.1109/INDIN51773.2022.9976159
M3 - Conference contribution
AN - SCOPUS:85145776728
T3 - IEEE International Conference on Industrial Informatics (INDIN)
SP - 27
EP - 32
BT - 2022 IEEE 20th International Conference on Industrial Informatics, INDIN 2022
PB - Institute of Electrical and Electronics Engineers Inc.
Y2 - 25 July 2022 through 28 July 2022
ER -