Acceleration‐Based Automated Vehicle Classification on Mobile Bridges

Mobile bridges have been used for a broad range of applications including military transportation or disaster restoration. Because mobile bridges are rapidly deployed under a wide variety of conditions, often remaining in place for just minutes to hours, and have irregular usage patterns, a detailed record of usage history is important for ensuring structural safety. To facilitate usage data collection in mobile bridges, a new acceleration‐based vehicle classification technique is proposed to automatically identify the class of each vehicle. Herein we present a new technique that is based on the premise that each class of vehicles produces distinctive dynamic patterns while crossing this mobile bridge, and those patterns can be extracted from the system's acceleration responses. Measured acceleration signals are converted to time–frequency images to extract two‐dimensional patterns. The Viola–Jones object detection algorithm is applied here to extract and classify those patterns. The effectiveness of the technique is investigated and demonstrated using laboratory and full‐scale mobile bridges by simulating realistic scenarios.