| Summary: | This work evaluates the pavement surface condition and detects the pavement subsurface delamination through vehicle noise collected by microphones mounted underneath a moving vehicle. Such measurements will include tire-generated sound, which carries much information about the road condition, as well as noise generated by the environment and vehicle. A careful frequency analysis of the vehicle noise is carried out to localize the frequency range for surface characteristics
and subsurface delamination respectively. The Principal Component Analysis (PCA) method is applied to differentiate important information about the road condition from noisy data contributions collected while the vehicle is moving. The analysis begins with acoustic pressure measurements made over constant and known road conditions. Fourier transforms are taken over various time windows and a PCA is performed over the resulting vectors, yielding a set of principal component vectors for
the road condition. The condition of each road section is characterized by a set of principal component vectors. Pavement macrotexture Mean Texture Depth (MTD) is predicted from the principal component vector and then projected to the overall pavement condition index (PCI) as well as the pavement friction and uniformity evaluation. Pavement subsurface delamination is also detected from the principal component vector. The significant accomplishments of this study are as follows: (1)
localize the frequency ranges of vehicle noise related to pavement surface and subsurface features respectively; (2) demonstrate the potential for PCA to reduce noise and illustrate the procedure to apply PCA in frequency domain for noise filtering; (3) find the relationship between MTD, sound pressure and driving speed; (4) optimize the microphone placement for data collection for the MTD prediction; (5) develop a calibration-free method for the MTD prediction with an error of 16%; (6)
use the predicted Equivalent MTD found by the PCA Energy Method to explore pavement friction prediction, construction non-segregation evaluation, and the contribution to PCI prediction from the acoustic aspect; (7) differentiate the stiffness of the top layer pavement within 10 cm of surface from tire/road noise, and (8) detect the subsurface delamination from the shift of frequency in peak sound pressure level.
|
|---|
