Human System Exploration for the AI-based flight deck decision support system IPAS

Kurzfassung

The integration of human users with AI-based systems is crucial for the development of AI assistance systems in safety-critical work environments. The German Aerospace Centre (DLR) is developing the Intelligent Pilot Advisory System (IPAS) to demonstrate and research the application of AI in flight decks. The development of the IPAS started with a number of high-level requirements and general ideas for use cases. In parallel with the system development, research focused on how to provide AI explanations tailored to the end user, in this case airline pilots. In order to make creative and effective use of the freedom given and to identify the requirements for explainable AI in the flight deck, the IPAS was developed and researched in an exploratory manner following the Human System Exploration (HSE) paradigm from initial vision to an executable demonstrator in a simulation environment. This paper outlines the reason for choosing the HSE paradigm, the modifications made and its practical application. It describes the development and research process, introduces the methods used, presents the results and discusses the value of this approach in terms of scientific research and the development of the IPAS.