Despite on-the-job assistance technology is getting popular in the Smart Operator 4.0 literature, non-routine work also requires highly-skilled and pre-trained workers to prevent serious errors and keep high efficiency and safety levels. The question that motivates this study is: 'is it possible to train industrial workers for what exactly is coming instead of preparing them for a large set of scenarios - even very unlikely ones?'. This work proposes: (i) a structured On-the-Job Training strategy for non-routine tasks, called 'training-on-the-go' or 'prescriptive training', according to which a Prescriptive Analytics module schedules the training sessions not long before the actual performance, but only when and if needed; (ii) a proof-of-concept of a game-based training system where virtual scenes and context of an industrial site are faithfully recreated thanks to digital twin data and models; (iii) the use of evolutionary-based fuzzy cognitive maps (E-FCM) for the extraction of the workers' implicit procedural knowledge and for the comparison of mental models of experienced vs. inexperienced workers to assess potential misconceptions or flaws in their decision-making process. This work contributes to the evolution of worker training paradigms and systems from the perspective of human-centric cyber-physical production systems and aims for current gaps in workforce training, i.e. poor timeliness and effectiveness, limited context and industrial information integration, and scarce focus on the experts' implicit knowledge. An application study with a non-routine task on an offshore oil platform demonstrates how the proposed system facilitates knowledge transfer, offers situational awareness and sustains the workforce competence development process.

From “prepare for the unknown” to “train for what's coming”: A digital twin-driven and cognitive training approach for the workforce of the future in smart factories

De Felice F.;Petrillo A.;
2023-01-01

Abstract

Despite on-the-job assistance technology is getting popular in the Smart Operator 4.0 literature, non-routine work also requires highly-skilled and pre-trained workers to prevent serious errors and keep high efficiency and safety levels. The question that motivates this study is: 'is it possible to train industrial workers for what exactly is coming instead of preparing them for a large set of scenarios - even very unlikely ones?'. This work proposes: (i) a structured On-the-Job Training strategy for non-routine tasks, called 'training-on-the-go' or 'prescriptive training', according to which a Prescriptive Analytics module schedules the training sessions not long before the actual performance, but only when and if needed; (ii) a proof-of-concept of a game-based training system where virtual scenes and context of an industrial site are faithfully recreated thanks to digital twin data and models; (iii) the use of evolutionary-based fuzzy cognitive maps (E-FCM) for the extraction of the workers' implicit procedural knowledge and for the comparison of mental models of experienced vs. inexperienced workers to assess potential misconceptions or flaws in their decision-making process. This work contributes to the evolution of worker training paradigms and systems from the perspective of human-centric cyber-physical production systems and aims for current gaps in workforce training, i.e. poor timeliness and effectiveness, limited context and industrial information integration, and scarce focus on the experts' implicit knowledge. An application study with a non-routine task on an offshore oil platform demonstrates how the proposed system facilitates knowledge transfer, offers situational awareness and sustains the workforce competence development process.
File in questo prodotto:
Non ci sono file associati a questo prodotto.

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11367/120336
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 9
  • ???jsp.display-item.citation.isi??? 6
social impact