Enhancing Time Series Analysis in Flight Testing With Real-Time Embedded AI

Abstract

The field of flight tests has traditionally relied on deterministic data, making the presence of on-board AI uncommon. However, as the number of measurement points in flight tests increases, Safran Data Systems (SDS) recognizes the need to address the growing data flow without scaling the entire acquisition chain. SDS has introduced embedded AI algorithms to reduce the load on the acquisition chain by filtering out nominal data, only keeping outliers. While initially used for monitoring electrical networks, this real-time time series analysis has vast potential. It could revolutionize pre-flight checks by enabling thousands of signals to be checked in real-time, accelerating the go/no-go decision process. SDS envisions using AI to unlock new levels of efficiency and data analysis in flight testing. The future of flight tests could involve intelligent systems that swiftly and accurately assess data, providing invaluable insights and expediting the flight testing process.