Research on Bio-inspired product design based on knowledge graph and semantic fusion diffusion model

Bio-inspired design (BID), as an innovative design methodology, holds a significant position in the field of design, providing essential means for exploring problem-solving approaches and creating aesthetically pleasing product appearances. However, two major challenges exist in the current domain of BID: accurately matching suitable biomimetic organisms and efficiently generating high-quality biomimetic design solutions. To address these issues, this study simulates the process of designers engaging in BID and proposes a novel approach to bio-inspired product design (BIPD) based on a knowledge graph and a semantic fusion diffusion model (SFDM). First, by thoroughly analyzing textual data describing the physiological and behavioral characteristics of various organisms, along with design requirements, we construct a biological knowledge ontology layer and a corresponding graph database tailored for BIPD. This allows for the accurate recommendation of biomimetic organisms by inputting different product design requirements into the BIPD knowledge graph. This solves the problem of accurate matching between the target domain and source domain in BID, ensuring consistency between the recommended biomimetic organisms and product design requirements. Second, by processing and filtering the images of recommended biomimetic organisms and product names, and inputting them into the SFDM, we enable the rapid generation of diverse biomimetic design solutions. Furthermore, to ensure the systematic nature and practical effectiveness of this method, we introduce a corresponding BIPD evaluation system. Evaluation metrics related to BIPD are refined to comprehensively assess and select the optimal biomimetic design solution. Finally, through three different case studies, we demonstrate the efficiency and potential of the proposed method in addressing the challenges of target and source domain matching and the rapid fusion generation of biomimetic design solutions in the field of BIPD. This validation confirms the feasibility and effectiveness of the method. Our approach injects new vitality into the innovation of BID, providing designers with intelligent design system tools for efficient design based on specific needs, thereby enhancing the design quality and productivity of BIPD.