Flight testing is an important phase during the development of an aircraft to validate the design. During flight, data is gathered, and design problems are identified and solved. The collected data are fundamental for the analysis and Aircraft are properly instrumented to generate large amounts of information. Such a huge amount of data needs to be properly evaluated and traditional methods and platforms are no more effective.
Flight testing is a significant cost contributor to the aircraft production life cycle and is still extensively deployed. Flight test programmes take several years, and more prototypes are built to reduce lead times. Strong adherence to rigour safety and certification requirements and generally unchanged circular advisories inhibits the potential improvement of flight test designs. Innovative algorithms and statistical estimation are not achieving its full potential in the industrialized flight-testing environment.
The methods in this proposal increase the quality and productivity of an experiment, leading to a required test point reduction or increased predictive capabilities. The purpose of this project is to define and implement a state-of-the-art platform able to support data analysis. This is achieved by adopting a complex hardware architecture to support big data analysis and implementing specific algorithms to support data correlation, time series management and statistical analysis.
Furthermore, to support flight test engineers, novel approaches based on machine learning are provided to support the technicians in detecting specific flight conditions. The same platform is also adapted to support the development of the Next Generation Civil Tilt Rotor Technology Demonstrator.