Cloud-Edge MLOps for Diagnostic Analytics and Anomaly Detection in Smart Office Digital Twins

Area of Science:

• Computer Science
• Artificial Intelligence
• Internet of Things

Background:

• Smart buildings generate vast data streams from distributed sensors, requiring intelligent management.
• Traditional cloud-centric solutions face limitations in latency and scalability for real-time monitoring.

Purpose of the Study:

• To present an edge-enabled Digital Twin framework for smart office environments.
• To integrate real-time data acquisition, distributed intelligence, and machine learning analytics.
• To overcome limitations of cloud-centric approaches for smart building monitoring.

Main Methods:

• A multi-layer architecture (sensor, cloud-edge intelligence, interaction) aligned with Digital Twin models.
• Deployment of autoencoder models for anomaly detection on time-series sensor data.
• Integration of Machine Learning Operations (MLOps) for continuous model lifecycle management.

Main Results:

• A prototype demonstrated effective end-to-end data flow and stable long-term operation in a real smart office.
• Reliable anomaly detection with low-latency response was achieved.
• The system enables real-time monitoring and data-driven analysis for improved situational awareness.

Conclusions:

• The proposed framework effectively integrates Digital Twin technology with edge AI and MLOps.
• This approach provides a scalable and efficient solution for smart building monitoring.
• The system enhances operational decision-making through real-time insights.