Enhanced processing of low signal-to-noise-ratio dynamic signals from pavement testing

Abstract

The dynamic responses of a pavement structure to traffic loading is typically collected to predict pavement performance or in-service life. Researchers have primarily focused on using the maximum deviation from the peak strain responses to the conjoint valley responses for investigating the structural response of the pavement. However, the existing batch-processing program is not capable of identifying peak and valley responses from the low signal-to-noise-ratio (SNR) electronic signals. This paper explored the use of low-pass filtering and wavelet filtering to de-noise the strain signals for data pre-processing, and proposed the adaptive maximum energy ratio method and standard deviation method to isolate the foreground loading areas from the background noise of the strain responses. Finally, an automated data processing software was developed to precisely identify the peak and valley responses. This software provides a significant advancement to the current state of the practice in dynamic signal processing for pavement evaluation and performance prediction applications.