Knowledge-based information extraction from datasheets of space parts

Murdaca, F. and Berquand, A. and Kumar, K. and Riccardi, A. and Soares, T. and Gerené, S. and Brauer, N. (2018) Knowledge-based information extraction from datasheets of space parts. In: 8th International Systems & Concurrent Engineering for Space Applications Conference, 2018-09-26 - 2018-09-28, University of Strathclyde.

[thumbnail of Murdaca-etal-SECESA-2018-Knowledge-based-information-extraction-from-datasheets]
Preview
 Text. Filename: Murdaca_etal_SECESA_2018_Knowledge_based_information_extraction_from_datasheets.pdf
Accepted Author Manuscript
Download (998kB)| Preview

Abstract

Selection of the right space parts is an essential step during the design of complex engineering systems and requires information that is typically embedded in unstructured documents like datasheets, Interface Control Documents (ICDs) and technical manuals. Satsearch (https://satsearch.co) aims to consolidate global space supply chain information within a single platform, by converting unstructured datasheets into machine-readable, human-readable, electronic datasheets (EDS). After satsearch’s initial, manual efforts at generating EDS from source datasheets for space parts, they realized that the process is not scalable. A possible solution is to employ knowledge base information extraction systems. The Design Engineering Assistant (DEA) team from the University of Strathclyde is currently working on the automation of the extraction of information from unstructured documents (e.g. textbooks, reports, datasheets, research papers, etc.) through the development of an expert system. This paper summarizes the approach and outcomes of a feasibility study for the DEA, assessing benefits and obstacles for the implementation of a fully-automated information extraction process, focusing at this stage only on datasheets for space parts for preliminary mission design.

ORCID iDs

Murdaca, F., Berquand, A., Kumar, K., Riccardi, A. ORCID logoORCID: https://orcid.org/0000-0001-5305-9450, Soares, T., Gerené, S. and Brauer, N.;
CORE (COnnecting REpositories)