Exécution de notre outil de synthèse sur un nouvel article

Résumé : Occupant-Centred Digital Twin (OCDT) is a novel approach that focuses on enhancing indoor comfort by dynamically responding to individual occupants’ preferences. The proposed OCDT model integrates four main modules: Data Acquisition, Digital Twin, Data Analysis, and Edge Intelligence. Environmental sensor data and occupant feedback—covering thermal, air quality, visual, acoustic, productivity, and cleanliness dimensions—are collected in real time, accounting for zone-based variations. The Digital Twin is modelled using Building Information Modelling (BIM) tools like Autodesk Revit and Tandem as a Digital Twin platform. Data Analysis includes preprocessing, feature engineering, and dimensionality reduction using PCA to compute an Occupants’ Comfort Score (OCS). Edge Intelligence leverages TinyML to deploy trained machine learning models—XGBoost, Random Forest Regressor, and Neural Network—on microcontrollers for real-time comfort prediction. Among them, XGBoost achieved the best performance with the lowest mean squared error (MSE = 0.0021). The model is validated through a case study in an office building in New Cairo, demonstrating its effectiveness in supporting proactive comfort management and optimised energy use.