From Idea to Engineering-Aware Product Design: A Case Study of Toy Car Development

This study presents an AI-based product design and analysis framework for generative manufacturing, integrating multimodal inputs with a human-in-the-loop system to overcome inefficiencies in conventional design-analysis iterations. Using a toy car as a case study, user inputs such as text prompts and sketches are transformed by AI models into images and 3D mesh representations. Engineering performance, including aerodynamic indicators, is rapidly predicted through AI-based surrogate models. When performance criteria are unmet or the user intends to modify the design, the system generates refined prompts and adjusted design proposals, enabling an iterative human-in-the-loop process. This approach allows the derivation of designs that satisfy both engineering performance and user intent, illustrating a paradigm in which AI effectively bridges conceptual design and engineering analysis.

Korean Society of Mechanical Engineers (KSME 2025)