AI Framework Automates Residential Floor Plan Compliance Checks

Governments in Australia have introduced policy reforms like SEPP65 and BADS to improve the quality of apartment design, focusing on health-related features such as daylight, ventilation, privacy, and space efficiency. However, ensuring compliance with these regulations is a manual, time-consuming process that struggles to scale for large numbers of apartments and evolving policies. Existing automated methods are often fragmented and limited to single units. To address these challenges, a conceptual framework is proposed for automated compliance checking in multi-apartment buildings. This framework leverages an LLM within a Rule Engine to translate textual building codes into actionable, explainable rules. Concurrently, a Data Extraction Engine processes floor plan images, segmenting them into elements like walls, rooms, and fixtures, and transforming them into a structured building graph that captures topological relationships. The Compliance Check Engine then evaluates this structured representation using the LLM-generated rules. This integrated approach promises a scalable, consistent, and transparent method for compliance assessment across different jurisdictions, thereby supporting efficient enforcement of design standards and promoting healthier urban development.