Application of Multisource Frequency Aggregation Method in Bridge Modal Frequency Identification within the Crowdsourcing Framework

Crowdsourcing-based indirect bridge health monitoring uses smartphone sensors to capture bridge vibration responses from volunteer vehicles. However, the coupling between vehicle-induced vibrations and bridge vibrations makes it challenging for traditional methods to accurately separate bridge modal frequencies. To address this issue, this study proposes the multisource frequency aggregation method, which integrates adaptive variational mode decomposition for noise suppression and extracts stable bridge frequency components based on similarity principles. The method utilizes the stability of bridge frequencies across multisource data while eliminating nonbridge frequencies that vary significantly due to vehicle parameters, enabling precise frequency identification. Numerical simulations and model experiments are conducted to validate the effectiveness of the proposed method. The results demonstrate that the method can accurately extract bridge modal frequencies without requiring prior knowledge of bridge or vehicle parameters. Moreover, the method exhibits strong robustness to variations in vehicle structure, speed, and traffic conditions. It provides a universal framework and a novel signal-processing approach for bridge health monitoring within the crowdsourcing framework.