Drive-by bridge damage detection using Mel-frequency cepstral coefficients and support vector machine

Bridge damage detection using vibration data has been confirmed as a promising approach. Compared to the traditional method that typically needs install sensors or systems directly on bridges, drive-by bridge gained increasing attention worldwide since it just one few instrumented passing vehicle. frequencies extracted from vehicle’s vibrations can be good references for detection. However, extant literature considered mainly low-frequency responses of vehicle, while high-frequency also contained bridge’s information were often ignored. To fill this gap, paper developed approach utilized both low and Mel-frequency cepstral coefficients (MFCCs) support vector machine (SVM) employed classify severity. Firstly, frequency are input features train SVM models identify condition. Then, reduce dimensions inputs improve training efficiency, projected Hertz scale into Mel scale, two means MFCCs used feed different models. A laboratory experiment with U-shaped continuous beam model car was verify effectiveness proposed method. Results showed contain much about conditions, could apparently computational efficiency. The errors when heavy within 5%.