Real-Life Investigations of Inverse Filtering for Frequency Identification of Bridges Using Smartphones in Passing Vehicles

This paper puts forward a real-life assessment of a novel inverse filtering methodology to extract bridge features from acceleration signals recorded on smartphones in the passing vehicles. The vibration of a moving vehicle is affected by various features, such as suspension and speed. This study focuses on filtering out these effects from the signals to extract bridge frequencies as the vehicle crosses the bridge. Hence, the spectrum of the vibration data recorded on the vehicle when moving off the bridge is employed to form an inverse filter which removes the vehicle-related frequency content. Since the speed of the vehicle is found to be one of the most effective factors in the filter design in our previous studies, a database of the off-bridge vibrations is built for different speeds. Later, when the same vehicle is moving on the bridge, the corresponding inverse filter is applied to the recorded on-bridge data to suppress the vehicle frequencies and amplify the bridge frequencies. All the required data are recorded using the built-in accelerometer and GPS sensor of the smartphone, eliminating the need for any extra instruments. In addition, this approach considers each data source separately and designs a unique filter for each data collection device within each vehicle, which makes it robust against device and vehicle features. As a result of the proposed methodology, it would be possible to monitor a large number of bridges using crowdsourced data collected from the smartphones in the vehicles. Such methodologies are expected to improve the sustainability and resiliency of our future infrastructure systems and future cities.