Accurate road damage classification based on real signal mother wavelet of acceleration signal

We propose a system called YKOB for road authorities to reduce road monitoring cost. It focuses on participatory sensing by cyclists wearing a smartphone. In this paper, we propose a method to classify road signal into three road conditions: positive step (PS), negative step (NS) and convex step (CS). In past studies, we tried to classify road signal with a feature-based algorithm. However accuracy is not sufficient, and it seems to be difficult to increase kinds of classifiable road conditions with this method. To resolve the problem, we develop a new road classification algorithm based on wavelet transform technique with Real-signal Mother Wavelet (RMW). We have developed two algorithms based on generalized and personalized RMW. 10 subjects participated to cycling experiments in the three road conditions. From the result, both algorithms were more accurate than feature-based algorithm.