Lee, Chiu-MingChiu-MingLeeSHANG-HSIEN HSIEH2026-01-152026-01-152026-0109265805https://www.scopus.com/record/display.uri?eid=2-s2.0-105023658228&origin=resultslisthttps://scholars.lib.ntu.edu.tw/handle/123456789/735368Smart construction sites are dynamic and risk-prone, requiring intelligent systems for continuous monitoring, contextual understanding, and autonomous hazard mitigation. Addressing the inherent limitations of centralized cloud-based systems, such as latency, scalability, and fragmented decision-making, remains a key challenge. This paper introduces an edge-enabled framework for real-time, context-aware hazard alerting, enabling efficient on-site data acquisition and predictive safety management. The framework shifts data processing from the cloud to the edge, leveraging distributed edge-enabled processing, autonomous event triggering, and queue-based data prioritization to achieve an unprecedented level of real-time responsiveness. The framework's effectiveness was validated through a simulation-based approach, utilizing both on-site sensor data (e.g., temperature, worker ID) and external datasets (e.g., PM2.5 indices, weather forecasts). This validates the system's ability to enhance predictive logic and adaptive decision-making. The proposed solution provides a scalable and robust foundation for next-generation management systems, significantly enhancing safety and operational efficiency in complex construction environments.falseContext-aware alertingEdge computingIoT integrationOpen dataSmart construction[SDGs]SDG13journal article10.1016/j.autcon.2025.1066422-s2.0-105023658228