Presentation
A RAG–LLM-Based Method for Automotive Human Factors Design Guidance
SessionThursday Poster Session
DescriptionAutomotive human factors knowledge is essential for early design decisions, yet relevant evidence is often dispersed across standards, design guidelines, regulatory resources, and research-based evidence. Designers must interpret these materials and translate them into actionable recommendations, which can limit timely use during conceptual design. This paper proposes a RAG–LLM-based method for automotive human factors design guidance. The method builds a task-oriented knowledge base from curated human factors materials, parses natural-language design inputs into structured intents, retrieves relevant evidence through semantic and terminology-based matching, and generates structured recommendations with source attribution. A prototype system demonstrated the workflow through a representative automotive HMI case, transforming an open-ended route-entry design task into evidence-grounded recommendations with inspectable sources. Preliminary user-oriented evaluation suggested usefulness for knowledge access, recommendation interpretation, and evidence review, indicating potential to support early-stage design reasoning.
Event Type
Poster
TimeThursday, October 22nd5:30pm - 6:30pm PDT
Location