Research on Cloud-Edge-Device Collaborative Computing Framework and Load Balancing Strategy for Highway Construction Safety Early Warning
Year: 2026
Type: conference-paper
Abstract: This paper proposes a cloud-edge-device collaborative framework and a DDQN-based dynamic load balancing strategy for highway construction safety early warning. To solve the high latency and network vulnerability of traditional cloud-centric systems in linear work zones with fluctuating workloads, a three-tier architecture (terminal sensing-edge computing-cloud intelligent control) is designed. Lightweight hazard detection models are deployed at the edge, and the DDQN algorithm schedules tasks based on node load, task urgency, and latency. Experimental results show: average warning latency <182 ms in peak conditions with autonomous operation during network outages; the load balancing strategy reduces imbalance by 46% and improves resource utilization to 92%; the detection model achieves 82.0% mAP with only 2.2 MB parameters. This study provides an effective intelligent safety monitoring solution for complex highway construction environments.
Authors: Zijie Zhang, Cong Chen
Institution: Xi'an Eurasia University
Language: English
Topic: IoT and Edge/Fog Computing
Subfield: Computer Networks and Communications
Field: Computer Science
Domain: Physical Sciences
Open Access status: hybrid
OpenAlex ID: W7213559136