This research could revolutionize Philips’ production by integrating advanced Artificial Intelligence and Machine Learning into key processes, shifting from manual oversight to real-time anomaly detection and predictive maintenance. This approach reduces downtime and defects. This work not only enhances production capabilities but could also set a new benchmark for technological advancement in the manufacturing industry, driving significant improvements in productivity and product quality.

ML-based anomaly detection for production line process control

Due to the complexity of the assembly process, anomalies can occur. These anomalies can be classified into three types: point anomaly (single data point that deviates from normal behavior, such as misplacement), contextual anomaly (data that is unusual only within a specific context), and collective anomaly (unusual pattern formed by a group of otherwise normal data points such as sensor fatigue). Based on the anomaly data, root cause analysis can be further performed to form hypothesis about the underlying causes for further investigation. Furthermore, this strategy can be applied to other implementations, increasing overall production efficiency.

Issue location detection in the production line using AI

Locating issues in the production line can be difficult, but this process can be facilitated by concentrating on how fixtures and jigs affect critical-to-quality (CTQ) metrics. By carefully evaluating these factors, it will be possible to create more intelligent process control strategies that take into account the impact of different variables. Innovative methods of tracking reliability can also be achieved with production pathfinder algorithms, which offer more insight into possible sources of variation and facilitate more proactive quality control. Overall, tackling issue location detection methodically can help to speed up production and ensure product quality.

Transfer function inference for production line process using AI

Overall, having an understanding of the interactions that happen in a complex system and process is key to the implementation of efficient control strategies and promoting continuous improvement. In this workpackage we will develop a framework for systematic analysis of data stemming from all sequential steps of a production line, to facilitate inference on interaction patterns between production steps.