Intelligent Data Understanding for Architecture Analysis of Entry, Descent, and Landing: Applications for Mars Missions

Abstract

Because Entry, Descent and Landing (EDL) system validations are limited in Earth environments, these technologies rely heavily on modeling and analysis tools to evaluate system performance and capture uncertainties, which determine the success of a mission. However, the current approach suffers from an important limitation. While the subject matter expert can leverage his or her knowledge and expertise with past systems to identify areas of risk and features of interest in the datasets available, the next generation of EDL systems may present unprecedented challenges that may be missed by the human. Landing humans on Mars, for example, will pose unprecedented challenges to EDL technologies driven by the need to land larger payload mass with extremely high reliability and safety requirements. The goal of this research is to advance the state of the art for offline Intelligent Data Understanding (IDU) technologies for Entry, Descent and Landing (EDL) by incorporating an intelligent assistant- called Daphne/EDL - that supports humans in architecting problems specific to the field of EDL. In this thesis, we describe a first prototype of the Daphne/EDL assistant in the context of three use cases to cover a range of representative problems in EDL architecture analysis and to show the capabilities of the assistant to support those use cases. Specifically, we demonstrate the baseline functionalities of the EDL assistant that include: 1) preliminary analysis capabilities of simulation datasets; 2) extraction of performance metrics from a historical database; and 3) automated generation of a scorecard. The scorecard contains metrics critical to assess EDL architecture performance. The assistant highlights metrics that fall out of spec of a design and communicates them to the subject-matter expert.