Effectivity evaluation of YKOB in actual city road
Ryuji Suzuki, Junji Takahashi, Yoshito Tobe · 2017
This paper deals with a road monitoring problem that aims to detect road anomalies with low costs. In order to cope with large area monitoring with low cost, we proposed a participatory sensing based monitoring system named Your Kinetic Observation Bike (YKOB) that only uses a smartphone and a bicycle previously. The monitoring system, YKOB, gathers acceleration signals recorded on smartphone that cyclists posses in a pocket or bag. Then the gathered data are analyzed through real mother wavelet (RMW) after independent component analysis (ICA) to predict the road conditions. Early type of YKOB system classifies the road conditions in to four types of step and the prediction accuracy overtook 0.8 in artificially set real environment. So as to evaluate the effectivity of the YKOB to real road condition, we gathered acceleration signals when a cyclist runs actual city roads and analyzed them with ICA and RMW based algorithm. In this paper we report the data gathering method in actual city roads, data analysis method, and their results after a brief introduction of the YKOB.