This use case focuses on end-to-end digitally driven luminaire manufacturing, including local fabrication of parts, intelligent production control, anomaly prediction, quality control, and digital-twin-based material and data flows. The platform combines human and machine processes with digital fabrication, computer vision, machine learning, AI-based scheduling, and real-time process monitoring. Its goal is to increase production efficiency, enable local-for-local manufacturing, reduce supply-chain dependencies, support easy assembly and disassembly, and move toward zero defect and zero waste manufacturing. The use case contributes to reshoring, reduced transport, higher resource efficiency, recycling of rejected 3D-printed parts, and adaptive production planning.
| # |
Technology Name |
Description |
| 1 |
3D metal printer development |
Development of a first-of-a-kind DED metal printer that fulfills the operational needs of product manufacturing for large-volume, lower-tech products, such as luminaires for Signify. |
| 2 |
Computer Vision |
Exploration and implementation of computer vision for the quantification of manual assembly, including process step compliance, throughput time, and related assembly process indicators. |
| 3 |
Intelligent Scheduler |
Improvement of scheduling and planning for order fulfillment in low-volume, high-variability 3D printing of polymer luminaires. |
| 4 |
FDM/FFF progress |
Increase the productivity of FDM/FFF additive manufacturing using polymer PC material. |
| KPI # |
Related Technology |
KPI Target |
| 1 |
FDM productivity |
Increase productivity, measured as printed or deposited material, by a factor of 2–3 compared to the baseline. |
| 2 |
New metal 3D printer |
Develop a first-of-a-kind printer, FUMO, that fulfils the requirements of consumer-like products in lighting. |
| 3 |
Intelligent scheduler |
Adapt order execution on an hourly level instead of a daily level. Automate the scheduling process toward human-less scheduling, with the purpose of minimizing waste and fulfilling order requests. |
| 4 |
Intelligent scheduler |
Develop a next-level scheduler that takes multiple anomaly types into account, as part of the scientific study at TU/e. |
| 5 |
Computer Vision |
Acquire quantified data from the assembly line for manual process execution. Automate this process and acquire more details and functionalities. |
| KPI # |
Baseline |
| 1 |
Productivity moved from x kg/h to 2x kg/h. The exact values are company-confidential and can be shared with reviewers. |
| 2 |
The baseline is a high-cost, above 500 kEuro, low-volume and slow DED 3D metal printer available on the market, which does not clearly match the business case for luminaire fabrication. |
| 3 |
Single-day, paper/whiteboard-based planning through stand-up meetings. |
| 4 |
Baseline input should be provided by the TU/e partner. |
| 5 |
Manual quality and execution monitoring, performed by a process line engineer using methods such as stopwatch-based timing. |
| KPI # |
Validation Method |
| 1 |
Experimentation in lab equipment, followed by verification in the factory after formal transfer of the new technology. |
| 2 |
Realization of a functional FUMO model, including process development, print efficiency evaluation, and cost modeling based on machine cost and cost per printed object. |
| 3 |
Check the increase of order fulfillment and the higher efficiency of the planner in pilot trials at the 3D printing hub in Turnhout, Belgium. |
| 4 |
To be defined by TU/e. |
| 5 |
Validate through an ongoing pilot trial in the Valladolid, Spain outdoor luminaire factory. |
- Industrial process phases made more effective and precise through integrated and tailored AI algorithms, methods, and tools.
- 30% improvement in time spent on quality control.
- Reduced re-configuration time of the AI-based system by approximately 30–50%.
- Improved overall productivity of the manufacturing factory by approximately 20%.
- Providing documented methods for dynamic application of tools to improve sustainability.
- Improved performance and faster time-to-market.
The use case contributes to the AIMS5.0 Objective 2 KPIs mainly through improved operational excellence in manufacturing. Additive manufacturing supports more sustainable and flexible local production, while the FDM/FFF productivity improvements and the new FUMO metal printer contribute to higher manufacturing productivity and more efficient fabrication of luminaire components. Computer vision supports the automation of quality control and manual assembly monitoring by collecting quantified process execution data. The intelligent scheduler contributes to more adaptive order execution by moving planning from a day-level process toward hour-level automated scheduling, reducing manual planning effort and improving responsiveness to order and production variability. Overall, the use case supports productivity improvement, quality-control automation, sustainable manufacturing methods, and faster adaptation of production planning.
| AI Toolbox Tool |
Use within the Use Case |
| Object detection tool |
Used to support computer-vision-based monitoring of manual assembly processes. It helped detect relevant objects, tools, components, or assembly states on the production line. |
| Object detection with YoloV8 |
Used for real-time object detection in the assembly-line monitoring setup, supporting automated quantification of process execution and throughput-related observations. |
| Image Classification Tool |
Used to classify assembly process states, quality-related visual conditions, or compliance-related categories from production-line image data. |
| Multi-Object Tracking Tool |
Used to track objects or process-relevant elements across video frames, supporting the measurement of manual assembly execution, process timing, and movement through the production line. |
| Time-Series Anomaly Detection Tool |
Used to support anomaly-aware production monitoring and scheduling by detecting unusual temporal patterns in production, process, or order-execution data. |
| Tabular Prediction Tool |
Used to support data-driven scheduling and planning decisions for low-volume, high-variability 3D printing production, based on structured production and order data. |