Noise-based pavement condition inspection method with dynamic time warping

Traditionally, pavement condition inspections have been dependent on visual-based information, but there are disadvantages that it requires a lot of manpower, equipment and budget. Especially, very expensive automated inspection vehicles are essential for network level pavement inspection. This is considered as a major obstacle to the introduction of the pavement management system of road agencies in poor management environment. This study proposes a method determining the pavement service level using the pavement-tire friction noise. Noise measurement specification is according to the ISO11819-2, and judging the service ratings for noise profiles is conducted by the K-nearest neighbor with Dynamic Time Warping (DTW), a pattern recognition algorithm categorized in the machine learning. This simple idea and technique could present a new paradigm that replaces traditional visual-based inspection methods.