When Construction Supply Chains Meet 6G: A Deep Neural Network-Based Real-Time Data Transmission Approach

Abstract

Traditional communication infrastructures often struggle to support the demands of real-time data exchange required for modern construction practices like building information modeling, drone monitoring, sensor networks, and automated equipment, leading to delays, cost overruns, and suboptimal resource allocation. This letter presents a deep neural network-based real-time dynamic selection (DRDS) algorithm for modern construction supply chains that leverages 6G network capabilities for ultrafast data transmission. The approach uses historical project data to train a deep neural network model that dynamically selects optimal priority rules for resource allocation and scheduling based on real-time project status. Experimental results demonstrate that DRDS outperforms existing methods, achieving 95.2% relative optimal solutions for large-scale projects while maintaining solution times under 1.12 s. When deployed on 6G networks, the algorithm achieves 0.23 ms transmission latency, 39.2% bandwidth utilization, and can support 12 580 sensor nodes per km².