생성 AI 활용 범죄예방디자인 적용 시각화 비교 개발 사례

초록

The importance of Crime Prevention Through Environmental Design (CPTED) has been continuously emphasized. However, the actual design phase largely relies on text-based manuals and case-oriented guidelines, presupposing a high level of professional understanding of environmental design and consequently exhibiting limitations such as strong expert dependency and reduced design efficiency. To complement these structural limitations, this study proposes 'AI-CPTED', a design support system that integrates multiple specialized AI agents with Retrieval-Augmented Generation (RAG) technology. Based on a knowledge base of domestic CPTED-related regulations and guidelines, the system is configured to analyze current images of a target space, automatically generate improved alternative images reflecting CPTED principles, and provide an integrated quantitative analysis report on the application results. Empirical analysis showed that the generated alternatives achieved an average context-aware compliance rate of 83.2% for CPTED requirements. These results confirm that the proposed system supports consistent application of CPTED guidelines and provides an objective basis for evaluating the generated design alternatives. Consequently, this study holds significance in demonstrating the potential of generative AI technology as a practical tool to improve the accessibility and efficiency of CPTED design while objectively assisting the decision-making of design practitioners.