Eliciting and Exploiting Procedural Knowledge in Industry 5.0

Informations projet

PERKS

N° de convention de subvention: 101120323

DOI

10.3030/101120323

Date de signature de la CE 3 Juillet 2023

Date de début 1 Octobre 2023

Date de fin 30 Septembre 2026

Financé au titre de

• Digital, Industry and Space

Coût total

€ 3 550 250,00

Contribution de l’UE

€ 2 908 625,00

Coordonné par

CEFRIEL SOCIETA CONSORTILE A RESPONSABILITA LIMITATA SOCIETA BENEFIT
Italy

Periodic Reporting for period 1 - PERKS (Eliciting and Exploiting Procedural Knowledge in Industry 5.0)

Période du rapport: 2023-10-01 au 2025-03-31

Résumé du contexte et des objectifs généraux du projet

The PERKS project addresses a key challenge in modern industry: managing procedural knowledge - the know-how required to perform specific tasks. Poor procedural knowledge management can lead to errors, compliance issues, and onboarding difficulties. This is especially critical in manufacturing, where digital support for procedures is often lacking.

PERKS aims to enhance how procedural knowledge is captured, shared, and applied using AI and data technologies. Rooted in Industry 5.0 principles, the project focuses on human-centric innovation, supporting workers with intelligent tools that improve safety, efficiency, and knowledge transfer. It promotes social sustainability (worker well-being), economic sustainability (process efficiency), and, to some extent, environmental sustainability (resource optimization).

The project integrates AI, data technologies, and a human-in-the-loop approach to ensure usability and trust. It proposes a digital infrastructure for procedural knowledge management, including interoperable tools, methodologies, and a shared ontology. The goal is to enable flexible, resilient, and sustainable industrial workflows across Europe.

Travail effectué depuis le début du projet jusqu’à la fin de la période considérée dans le rapport et principaux résultats atteints jusqu’à présent

During the first 18 months, PERKS carried out an extensive requirements analysis across three industrial use cases and defined a reference architecture for its digital infrastructure. Based on this foundation, the project developed and released the first version of seven technical solutions:
- Automatic Knowledge Extraction
- Manual and Semi-Automatic Knowledge Collection
- Knowledge Validation with Human-in-the-loop
Procedural Knowledge Auditing
- Procedural Knowledge Management System
- Conversational Chatbot
- Conversational FAQ Generation

These solutions were integrated, deployed and evaluated across three diverse industrial use cases, each addressing a distinct procedural knowledge challenge. At Beko Europe, the focus was on improving Lock-out-Tag-out (LOTO) procedures—critical for ensuring machinery is safely deactivated before maintenance. PERKS helped formalize undocumented procedures, integrate worker input, and deploy a conversational AI assistant to guide operators step-by-step, enhancing both safety and confidence. At Fagor Automation, PERKS streamlined the complex commissioning process for CNC systems by digitizing expert knowledge and enabling junior technicians to access it through an AI-powered chatbot, reducing onboarding time and minimizing errors. At Siemens, the project supported microgrid management by capturing the rationale behind energy distribution rules and making it accessible via a chatbot interface. This not only empowered grid operators with real-time explanations but in perspective can also engage end users in adopting energy-efficient behaviors, contributing to smarter and more sustainable energy use.

The evaluation confirmed the relevance of the project’s objectives and its applicability across diverse scenarios. Feedback from users and external stakeholders highlighted the timeliness of PERKS and its potential for broader adoption. The Procedural Knowledge Ontology (PKO), which underpins all solutions, received strong interest from the scientific and technical community, supported by two publications. The project also identified additional use cases, including applications in digital twins and robotics.

Progrès au-delà de l’état des connaissances et impact potentiel prévu (y compris l’impact socio-économique et les conséquences sociétales plus larges du projet jusqu’à présent)

PERKS advances the state of the art by combining cutting-edge technologies - AI, generative AI, conversational agents, ontologies, and knowledge graphs - to support the full lifecycle of procedural knowledge. Its solutions enable automatic and semi-automatic knowledge capture, human-in-the-loop validation, and conversational access to procedures through retrieval-augmented generation (RAG) techniques.

A key innovation lies in the methodological integration of human feedback at every stage - from knowledge extraction and error handling to procedure approval and auditing. This ensures that AI tools remain usable, trustworthy, and aligned with real-world needs. The Procedural Knowledge Ontology (PKO) provides a semantic backbone that ensures consistency and interoperability across tools. PERKS also brings digital innovation to traditionally non-digital industrial contexts, enabling safer, more efficient, and standardized operations.

High-level view of the PERKS project