Detection of Road Damage using Signals of Smartphone-Embedded Accelerometer while Cycling

Recently, bicycle accidents have been increasing. One of the causes is poor road maintenance because of its high costs. The purpose of this study is the reduction of maintenance cost. We propose a method to detect road damage from cyclists pedaling motion. First, cyclists wear a smartphone in their pants' side pocket and collect acceleration signals with embedded accelerometer. Second, the collected signals are separated into road signal and pedaling signal by applying Independent Component Analysis (ICA). Finally, road condition is classified depending on road signal's feature. We carried out experiments on three road conditions: flat, positive step damage, and negative step damage. Using proposed method, road condition's classification overall accuracy was 70%.