Host: The Japan Society of Mechanical Engineers
Name : [in Japanese]
Date : December 01, 2021 - December 03, 2021
This report describes the results of examining a method for detecting track fault from train running sound data. Specifically, we installed a microphone under the floor of the car to record the running sound of the train. Then, using this running sound data, we were able to construct an algorithm that can discriminate between a normal state and an abnormal state of the track. However, this is the case for some tracks, and it turns out that there are some challenges to using it on various tracks.
