YAPAY ZEKÂ DESTEKLİ BIM ANALİTİĞİ VE DİJİTAL İKİZ GELİŞTİRMEYE YÖNELİK ENTEGRE HESAPLAMALI YAKLAŞIM

Read full text Summarise with VIRAT
Authors: Nilsu Şahin
Cited by: 0

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

Building Information Modeling (BIM) integrates geometric representations with structured spatial and functional data, enabling reuse in interactive and operational digital twin applications beyond design and construction. Existing BIM models can be incorporated into computational pipelines to support indoor navigation and emergency evacuation in complex facilities. This thesis introduces an integrated framework for AI-assisted BIM analytics and digital twin development. It shows how Revit-based BIM data can be transformed into a unified pipeline for spatial analysis, navigation, and human-centred interaction within a Unity-based environment. The workflow separates geometric representations from structured spatial data to enable real-time indoor pathfinding with the A* algorithm, while accounting for key physical constraints relevant to evacuation. A large language model (LLM) serves as an interaction layer, translating computed routes into natural-language navigation instructions and supporting context-aware user queries. The system provides real-time visualization and interaction via a Unity-based digital twin, enabling interactive testing and demonstrations. The framework supports both algorithmic and human-centred evaluation. Thirty participants completed user evaluations using the System Usability Scale (SUS) and the Usefulness, Satisfaction, and Ease of Use (USE) Questionnaire. Overall, the proposed approach offers a scalable solution for integrating BIM-based navigation, AI-driven interaction, and immersive digital twins for indoor and emergency-oriented applications.

Fields of Study & Subjects

BIM and Construction Integration 3D Modeling in Geospatial Applications Architecture and Computational Design

Publication details

Journal / source
OpenMETU (Middle East Technical University)
Publication date
16 March 2026
Language
EN
Document type
Thesis