Exploring Crowdsensing Potentials for Structural Health Monitoring Applications: Challenges and Opportunities

The use of devices equipped with sensing capabilities (ranging from smartphones to smartwatches) has become commonplace in our daily lives. These devices can measure various parameters such as acceleration and position, and have the potential to be leveraged for purposes beyond their intended use. This is the main goal of crowdsensing techniques. However, if on the one hand crowdsensing is spreading for its great appeal, on the other hand sometimes the importance of key concepts related to the robustness of the measurement data collected are underestimated. This is particularly true when crowdsensing techniques target structural dynamic problems. Indeed, the accurate extraction of dynamic information beyond the resonant frequencies of the structure is still an open issue and requires tackling the problem considering metrological aspects. Including effects like different sensitivities, sampling frequencies of the devices, as well as all the other interfering inputs that might affect a structural dynamic measurement, will provide the basis for extracting enriched structural data like mode shapes and damping. This may pave the way to a paradigm shift in monitoring civil infrastructures, as fixed installations could be correlated, compensated, or even substituted by mobile solutions. To prove these potentials, this paper discusses the main challenges arising in this application field of crowdsensing, such as the need of synchronizing the devices if targeting operational deflection shapes (ODSs) reconstructions, the lower sensitivity of sensors incorporated in mobile/wearable devices (typically based on micro electro-mechanical system - MEMS - technology), etc. These issues are demonstrated and tackled by performing a three-step analysis. At first, the metrological characterization of the selected hardware is carried out. In the second step, a first experiment is conducted on a laboratory case study, to verify whether the proposed system can capture the dynamic response of the structure. The final step consists of an experimental campaign conducted on a pedestrian bridge. The results obtained clearly demonstrate the need for a multidisciplinary approach which includes metrology science when targeting crowdsensing for the assessment of the dynamic behavior of a structure.