Leveraging AI-driven requirements for SysML modeling of the IoBT: A comprehensive investigation

Abstract

Designing secure and adaptable systems for the Internet of Battlefield Things (IoBT) presents substantial challenges due to the complexity, variability, and high-risk operational environments inherent to modern military engagements. Deriving accurate and traceable system requirements is particularly difficult given the dynamic threat landscape and the heterogeneous nature of IoBT components. This study conducts a systematic literature review to investigate how artificial intelligence (AI), particularly Large Language Models (LLMs), can augment the derivation of cybersecurity requirements and support their structured representation through Model-Based Systems Engineering (MBSE) and Systems Modeling Language (SysML) frameworks. By synthesizing findings across current research, this work identifies key limitations in traditional manual processes. It demonstrates how AI-driven automation improves the efficiency, precision, and adaptability of security engineering in military systems. The review also highlights critical gaps in AI-to-SysML translation, traceability mechanisms, and real-time threat responsiveness. This research contributes a foundational understanding of how LLMs can be operationalized for secure systems modeling in IoBT contexts, offering a path forward for scalable, resilient, and compliant cybersecurity architectures in defense applications.