- Starion Italia is supporting FadeOut Software to improve the European Space Agency’s (ESA’s) REX (Return of Experience) Digital Dashboard.
- Starion’s team will develop a text mining tool to broaden the range of data available to ESA engineers for identification and analysis of mission anomalies.
- A security risk assessment and gap analysis will also be carried out by Starion.
Starion Italia is working with FadeOut Software to improve ESA’s ability to identify and analyse anomalies as part of the reliability, availability, maintainability and safety (RAMS) section’s activities (TEC-QPR). Starion’s artificial intelligence (AI) experts will develop a text mining solution to automate the extraction of anomaly report data for use in the REX Digital Dashboard. Alongside this, Starion will use its extensive experience of security for space missions and applications to carry out a risk assessment to ensure the enhanced dashboard meets ESA’s security requirements.
The REX Dashboard provides ESA engineers with insights on dependability performances and an effective tool to improve prior and future reliability assessments of ESA’s missions by aggregating and visualising performance and other data. Currently, this relies on manual analysis of anomaly reports, which is time-consuming and limits the number of missions that can be integrated into the tool. To improve this and facilitate wider adoption of the dashboard, Starion’s team will design a text mining tool that uses AI techniques, including natural language processing (NLP), to extract this data directly from ESA’s Anomaly Report Tracking System and other sources.
Andrea Cavallini, AI/ML Competence Lead at Starion, said: “NLP and large language models [LLM] have now been developed to the extent that they can be used effectively to characterise data patterns that previously wouldn’t have been found. We will take an iterative approach to identify the best possible configuration of NLP and LLM for this project and develop the text mining tool using machine learning operations [MLOps] practices. This will help improve the reliability, scalability and maintainability of the text mining tool, and enable continuous improvement based on real-time data and feedback.”
Starion’s other contribution to this project will be to define and execute a security risk assessment and then identify and analyse any gaps between ESA’s System Security Requirements Specification and current security controls in order to ensure compliance and provide the basis for ESA security certification.
