About
The Project
MANUFACTOR project brings together 14 partners from 8 European countries for an ambitious 36-months. Together, we are turning the vision of Industry 5.0 into a reality by validating human-centric augmentation in the factories of the future.
By bridging the gap between technology and human potential, we are developing solutions that deliver solutions that reduce physical strain of workers, lower mental stress, and boost confidence by over 20%. Production and Plant Managers can expect to save up to €1M annually per medium-sized plant. This is achieved by cutting employee onboarding times by 35%, reducing manufacturing errors by 20%, and significantly lowering injury-related absences.
© Fill Gesellschaft m.b.H.
The Challenge
An ageing workforce
Experienced professionals are retiring, taking decades of valuable, practical “know-how” with them.
A changing generation
Newcomers face steep learning curves, intense mental demands, and repetitive physical strain.
Traditional automation often ignores the human element, leading to high onboarding times, workplace strain, and skills gaps.
The Manufactor Solution
© Fill Gesellschaft m.b.H.
Instead of equipping workers with heavy, uncomfortable wearable machinery, MANUFACTOR uses an intelligent, subtle network of sensors and software that adapts the factory floor to human comfort. Our system operates across four integrated technological layers.
The OMNI-SENSE Suite
Fixed cameras track body posture and worker proximity to machinery to check safety boundaries without recording personal identities or facial features.
Comfortable bracelets monitor heart rate variability, skin temperature, and stress indicators to map physical fatigue.
Cognitive monitoring features through binocular eye tracking, to automatically trigger workplace alerts and adaptive interventions when thresholds are exceeded. It will be possible to monitor the gaze patterns thanks to eye-tracking technology, and to show content on the XR display to visually guide the workers during certain tasks.
Sensor-equipped gloves and miniaturised tools extend human sensory capabilities during complex quality assembly checks.
Human Digital Twins
An operator’s physical and cognitive state is securely tracked in real time by an artificial intelligence model known as a Human Digital Twin (HDT). The HDT acts as a digital safety sandbox. Before any automated system sends an instruction to a worker or modifies a machine, it runs a quick simulation inside the twin to ensure the suggestion is safe and does not cause physical or mental overload.
Knowledge graphs
To ensure senior knowledge is never lost, MANUFACTOR uses automated software pipelines to digitise industrial manuals, standard operating procedures, and incident logs. We conduct interactive chat sessions with veteran technicians to extract their unspoken expertise, blending it all into a single, unified “digital memory” of factory knowledge.
Smart AI Augmentation Agent
Once validated, our multi-agent platform coordinates actions directly on the shop floor via specialized AI assistants:
This agent processes raw 3D skeletal data to automatically calculate real-time strain scores. If an operator experiences extreme strain, the agent redistributes tasks across the team or suggests micro-breaks via smart glasses interfaces.
The system monitors physical effort and the predictive models evaluate if an operator is falling behind a fast pace or facing high stress. It automatically sends commands to slow down a conveyor or adjust a collaborative robot's trajectory.
An on-premise, secure, large language model that acts as a real-time conversational assistant. To minimize AI hallucinations and ensure traceability, the model is grounded directly in company-specific knowledge graphs enriched with structured factory text and raw operator expertise. Outputs deliver hands-free instructional guidance through XR glasses and web dashboards.
Validation in real-world industrial pilots
To prove that the MANUFACTOR system works seamlessly across completely different manufacturing sectors, the platform is refined in two dedicated learning factories (at TUDa in Germany and TFCC in Greece) and validated in three distinct industrial environments:
High-Precision Equipment
FILL - Austria
Streamlining operator onboarding and reducing trainer supervision load by up to 40% through interactive XR training.
Automotive Assembly
Volkswagen Autoeuropa - Portugal
Improving physical comfort and reducing task uncertainty during manual alignment and quality checks by using miniaturised sensors and virtual gauges integrated into the XR glasses.
High-Voltage Transformers
Hitachi Energy - Sweden
This agent processes raw 3D skeletal data to automatically calculate real-time strain scores. If an operator experiences extreme strain, the agent redistributes tasks across the team or suggests micro-breaks via smart glasses interfaces.
Project Impact
Shorter onboarding, safer conditions, and higher confidence.
>35% reduction in new employee training times and >50% increase in worker satisfaction.
Advanced trustworthy AI, explainability, and multi-modal edge data fusion.
≤7% hallucination rate in FactoryGPT and ≥90% explainability coverage on all AI decisions.
Decreased production costs and fewer factory mistakes.
≥20% reduction in human-related errors and structural quality defects.
Lower raw material and energy waste from avoided errors.
Millions of Euros saved annually across scaled industrial implementations.
Ethics and Transparency
The MANUFACTOR project is built upon a fundamental “ethics-by-design” framework. Its core philosophy dictates that artificial intelligence, and advanced technologies must be deployed to support humans on the shop floor, never to replace human workers or automate their value.
The consortium adheres to the Horizon Europe ‘ethics by design’ principles and the harmonised rules of the EU AI Act, the Europe Regulation (Article 19), and the European Charter of Fundamental Rights.
MANUFACTOR has received the commitment of the FILL, HE, and VWAE Ethics and Legal Board departments for consultation and interaction. Composed of independent experts in AI safety, human participation, and data protection, this board enforces strict “go/no-go” gates before any technical installation.
The HE workers’ union will act as an advisor during design and implementation, guiding inclusivity, trustworthiness, and GDPR compliance.
All research participation involving workers, managers, and trainees is strictly voluntary and subject to explicit, informed consent.
MANUFACTOR ergonomic and biometric signals (such as heart rate, gaze, and posture data) operate in strict compliance with the GDPR:
- Data is pseudonymised, or anonymised directly on-device.
- No personal identifiers are stored with raw sensor data.
- No sensitive personal attributes (e.g., ethnicity, religion) are processed, and all data remains securely within the EU/Associated Countries.
For every pilot site (Sweden, Austria, Germany, Portugal, and Greece), MANUFACTOR maintains a living compliance dossier and issues strict readiness notes prior to any live demonstration.