A2.2 - Towards Edge AI in Flight Test Data Acquisition: A Proof of Concept for Anomaly Detection in Aircraft Electrical Networks
- Event
- ETTC 2024 - European Test and Telemetry Conference
2024-06-11 - 2024-06-13
Nuremberg - Chapter
- ML & AI Session I
- Author(s)
- R. Pelluault, G. Guerrero, S. Sivakumaran - Safran Data Systems, Les Ulis (France)
- Pages
- 32 - 39
- DOI
- 10.5162/ETTC2024/A2.2
- ISBN
- 978-3-910600-02-7
- Price
- free
Abstract
In recent years, the proliferation of artificial intelligence (AI) technologies and their associated use cases has been observed across various industrial and consumer domains. Within the field of instrumentation, the longstanding trend of acquiring and digitizing physical measurements captured by sensors has been further bolstered by the rise of edge computing, enabling early data-to-information transformation. Consequently, the emergence of edge AI becomes a natural extension of this evolution. However, the deployment of edge AI in the realm of embedded instrumentation introduces new possibilities while also challenging fundamental aspects of the field, including metrology, determinism, miniaturization, and power saving. In light of these considerations, Safran Data Systems has developed a Proof of Concept aimed at integrating an edge AI capability into a flight test data acquisition unit. The primary objective is to leverage this technology for anomaly detection in time series data pertaining to aircraft electrical networks monitoring. This paper presents the technological approach employed, highlighting its unique features and advantages. Furthermore, the outcomes and results obtained through experimentation are shared, shedding light on the effectiveness and potential of the implemented edge AI solution. Additionally, the paper explores future use cases and potential enhancements, paving the way for continued research and development in this area.