AI-Driven Test Flow Generation from Semi-formal Functional Safety Requirements

AI-Driven Test Flow Generation from Semi-formal Functional Safety Requirements

연구 분야: Software Development

학회: European Conference on Software Process Improvement

The advent of Artificial Intelligence (AI) has revolutionized productivity and quality across industries. Large Language Models (LLMs), such as Generative Pretrained Transformer (GPT) by OpenAI, exhibit remarkable natural language understanding and generation capabilities. In this study, we explore the application of LLMs for generating test flows from semi-formal functional safety requirements in the context of ISO 26262:2018. Our methodology involves prompting the LLM to understand and interpret semi-formal safety requirements to generate coherent and contextually relevant test flows specifically tailored for functional testing. These flows serve as valuable resources for test planning, execution, and verification. The first experimental results demonstrate the effectiveness of our approach. We also discuss challenges, limitations, and potential enhancements. Leveraging LLMs for test flow generation offers a promising avenue to enhance testing efficiency and ensure robust system behavior.