An Immersive Digital Twin Framework for Intelligent Spatially Grounded Anomaly Detection in Smart Homes

This research outlines a grounded digital twin for spatial anomaly diagnosis within smart homes.Unlike previous research where IoT sensing, anomaly detection and Virtual Reality (VR) visualization are treated as standalone, this study suggests a continuous workflow that spatianlly anchor anomaly detection within a VR digital twin.This work combines diverse IoT sensors in a VR digital twin of inhabited residences and an anomaly detection model (Isolation Forest and One-Class SVM) to interpret anomalies as room-level fault events rather than isolated time-series anomalies.An IoT sensor data are streams to firebase database where anomalies are detected with a noleak, strict time series, and spatial events are placed within 3D interactive models and mobile instance updates.Geographical resolution of detected anomalies facilitates the transformation of anomaly detection into diagnosis, streamlining human interpretability and accelerating fault location.These processes were examined in a vetted user study with 15 subjects under three diagnosis configurations: on-site, paper plans, and the proposed VR-AI digital twin.Results confirm the effectiveness of the proposed model, which achieved a mean diagnostic score of 8.1 out of 10, compared to 6.3 out of 10 and 5.81 out of 10 through on-site diagnosis and paper-based diagnosis, respectively.The model, on the other hand, achieved detection accuracies of 93% and 87% for water leak events and gas leak events, respectively, with improvements of 28.6% and 39.6% compared to physical and paper-based diagnosis, respectively.The system also achieved an average processing speed of 862 ms to provide instantaneous feedback.These findings demonstrate real improvment of fused unsupervised AI and VR-based digital twins for spatially anchored anomaly detection, yielding efficiency, speed, and reliability in fault detection compared to existing smart home maintenance systems.