2025
Moreira, Guilherme Micheli Bedini
Use of Large Language Models to Support Aerospace Defense Systems Engineering PhD Thesis
2025.
Links | BibTeX | Tags: Artificial Inteligence, Requirements, Systems Engineering
@phdthesis{moreira2025,
title = {Use of Large Language Models to Support Aerospace Defense Systems Engineering},
author = {Guilherme Micheli Bedini Moreira},
url = {https://cscerqueira.com.br/wp-content/uploads/2026/06/2025-Moreira.pdf},
year = {2025},
date = {2025-12-31},
urldate = {2025-12-31},
keywords = {Artificial Inteligence, Requirements, Systems Engineering},
pubstate = {published},
tppubtype = {phdthesis}
}
2024
de Mello Barbosa, Leonardo
2024.
Abstract | Links | BibTeX | Tags: Requirements
@phdthesis{nokey,
title = {NATURAL LANGUAGE STRUCTURING FOR DOCUMENTATION OF SYSTEMS REQUIREMENTS AND MILITARY EMPLOYMENT MATERIALS IN THE PORTUGUESE LANGUAGE},
author = {Leonardo de Mello Barbosa},
url = {https://cscerqueira.com.br/wp-content/uploads/2025/02/2024-Leonardo.pdf},
year = {2024},
date = {2024-12-31},
urldate = {2024-12-31},
abstract = {The thesis outlines a structured natural language for the documentation of requirements for military systems and employment materials, in the Portuguese language, for use in the Brazilian Army. Following the Design Science Research paradigm, the problem was precisely defined and justified in terms of its academic, industrial, and governmental relevance. Awareness of the problem arose through the exposition of the theoretical foundations of systems engineering and requirements engineering, and through the achievement of an integrative literature review. After establishing its objectives, the artifact was constructed using the design catalog of solutions method, resulting in a structure made up of 69 individual blocks. The artifact was artificially evaluated using a mathematical-linguistic model, based on dependency grammar. By processing 696 sentences, it was statistically proven that the use of the proposed structure is capable of reducing the textual complexity of requirements in three metrics relevant to the model. A naturalistic evaluation, conducted through the analysis of the perception of a group of typical users, demonstrated that it meets goals related to military use, sentence writing improvement, identification of structuring terms, and the Technology Acceptance Model. Evidence was found that, even though the ease of learning is uncertain, potential users are willing to employ the artifact in their professional daily routine.},
keywords = {Requirements},
pubstate = {published},
tppubtype = {phdthesis}
}
The thesis outlines a structured natural language for the documentation of requirements for military systems and employment materials, in the Portuguese language, for use in the Brazilian Army. Following the Design Science Research paradigm, the problem was precisely defined and justified in terms of its academic, industrial, and governmental relevance. Awareness of the problem arose through the exposition of the theoretical foundations of systems engineering and requirements engineering, and through the achievement of an integrative literature review. After establishing its objectives, the artifact was constructed using the design catalog of solutions method, resulting in a structure made up of 69 individual blocks. The artifact was artificially evaluated using a mathematical-linguistic model, based on dependency grammar. By processing 696 sentences, it was statistically proven that the use of the proposed structure is capable of reducing the textual complexity of requirements in three metrics relevant to the model. A naturalistic evaluation, conducted through the analysis of the perception of a group of typical users, demonstrated that it meets goals related to military use, sentence writing improvement, identification of structuring terms, and the Technology Acceptance Model. Evidence was found that, even though the ease of learning is uncertain, potential users are willing to employ the artifact in their professional daily routine.