Modular Sensor Platform for Real-Time Load Analysis of Infrastructure

Abstract

Monitoring large-scale infrastructure requires the integration of diverse sensors at various locations throughout the structure. This generates large data volumes whose secure transmission and analysis must be guaranteed. This work presents the development of a hierarchical, modular edge platform for load analysis of civil infrastructure. It comprises a variable number of Sensor-EDGE units with direct sensor interfaces and a central node called the EDGE-Gateway that establishes, among other functions, the connection to the cloud. Particular emphasis is placed on edge electronics that enable resource-efficient data preprocessing through FPGA-based hardware acceleration. As a use case, high-frequency vibration data acquisition and evaluation is examined, supplemented by camera-based object detection and environmental data. The vibration sensors utilize inexpensive commercially available MEMS accelerometers embedded directly in the road surface. The vibration data are evaluated primarily through classical signal processing, including integral energy content, maximum values, and frequency spectra. Crossing events are classified and compared for vehicles traveling at similar speeds for example, by overlaying frequency spectra from multiple passages of the same vehicle. The platform’s results are qualitatively compared with drive-over tests to assess system suitability. Measurement accuracy is estimated in the context of sensor noise. This road use case demonstrates the feasibility of the approach and positions the platform as a promising, cost-effective alternative to established systems such as geophones.