Development of a Low-Cost Wireless Sensor Network for AI-Based SHM in Steel Truss Bridges

Ensuring the structural integrity of steel truss bridges is critical for maintaining transportation safety and urban infrastructure resilience. This study presents the development and field validation of a low-cost wireless sensor network (WSN) integrated with artificial intelligence (AI) techniques for real-time structural health monitoring (SHM) of steel truss bridges. The proposed system utilizes a network of ESP۳۲-based sensor nodes equipped with low-power MEMS accelerometers and strain gauges, which communicate via the LoRa protocol to a centralized edge-processing unit. To reduce both deployment costs and energy consumption, custom-designed printed circuit boards (PCBs) and solar-powered modules were implemented. Real-time vibration and strain data were collected from a ۵۰-meter span steel truss bridge located in central Tehran, Iran, under varying load conditions and ambient noise levels. A deep learning model based on convolutional neural networks (CNNs) was trained using datasets encompassing both undamaged and artificially induced damage states, enabling the identification of early-stage structural anomalies. The model achieved an overall classification accuracy of ۹۴.۶%, demonstrating resilience against environmental noise and structural variability. These results validate the feasibility of combining cost-effective WSN hardware with AI-driven analytics to deliver scalable, autonomous SHM solutions for aging bridge infrastructure in both urban and developing regions.

Keywords:

Wireless Sensor Networks (WSN) , Structural Health Monitoring (SHM) , Steel Truss Bridges , Artificial Intelligence (AI) , Convolutional Neural Networks (CNN) , LoRa Communication , Vibration Analysis , Low-Cost Monitoring Systems , Real-Time Damage Detection , Edge Computing