Automated Reporting for Flight Tests: Enhancing Efficiency with AI and Digital Twin Technology

Abstract

Modern flight tests generate vast amounts of data at a pace that surpasses human interpretability. As a consequence, the need for structured reports addressing various stakeholders, ranging from sustainability and safety to management is rapidly increasing.

An essential part of every flight test is the creation of a flight test report. Although all necessary information is available digitally, flight test engineers must manually compile and verify it, a process that can take several days or even weeks depending on the complexity. In an effort to reduce unnecessary manual labor, this work investigates the automated creation of parts of such reports. Initial executive reports are created using sensor data, flight documentation, and other external data sources such as ADS-B traffic and METAR, which are analyzed with traditional as well as AI-powered tools. The contents comprise information about the time and date at which the flight took place, the flight trajectory, flown maneuvers, fuel consumption, METAR reports at departure and arrival, specific sensor values, measurement volume verification and the presence of nearby traffic. The report can either be created automatically in less then a minute, or customized individually on a website and only requiring quick human verification, leading to a significant decrease of manual labor.