XIA Ye, SHEN Zhou-hui, SHU Jiang-peng, SUN Li-min
Structural Health Monitoring (SHM) technology has become a key tool for ensuring the operational safety of bridges and supporting maintenance decision-making. However, the conventional bridge SHM paradigm relies on high-cost commercial sensors, closed data acquisition devices, and centralized analysis, resulting in high costs, limited scalability, and data transmission burdens. In this context, edge intelligence, with on-site information extraction and rapid response as its core, provides a new system-level pathway for bridge SHM to alleviate constraints related to cost, data transmission, and scalability. Low-cost sensors and embedded platforms respectively constitute the front-end sensing foundation and edge computing carrier of this paradigm, jointly promoting the transition of bridge SHM from the conventional paradigm characterized by high-cost device dependence, centralized raw data uploading, and cloud-diagnosis-oriented analysis toward a lightweight edge-intelligent paradigm characterized by low-cost dense sensing, scalable and controllable deployment, and edge-side information extraction. Following this main thread, this paper systematically reviews edge-intelligence-based lightweight bridge health monitoring. Firstly, the research status and paradigm evolution of lightweight bridge SHM are summarized. Secondly, from the sensing layer, the applications and technical boundaries of three types of low-cost sensors in bridge SHM are reviewed, including inertial and mechanical sensors, environmental and acoustic sensors, and optical imaging sensors. The common error sources and error calibration strategies of low-cost sensors are also summarized. Thirdly, from the transmission and analysis layer, the hardware performance of embedded platforms is systematically organized, based on which an in-depth analysis is provided of three typical architectures in lightweight bridge SHM systems: edge acquisition nodes, edge gateway nodes, and edge computing nodes. Subsequently, the system characteristics of lightweight bridge SHM is comprehensively reviewed from three dimensions: communication mechanisms, deployment architectures, and application scenarios. Finally, current technical bottlenecks are summarized, and future development trends are discussed in terms of multimodal sensing, edge-adaptive intelligence, and intelligent maintenance decision-making.