A Differential Sensor for Reliable Localization of Potential Cracks in Railway Tracks

This research paper and the resultant prototype are the initial outcome of the work of the authors to apply emerging learning based techniques for identification of rail track anomalies. As a divergence from the existing studies that used sonar or mechanical signals in identification of cracks that too mostly applied on robots or autonomous vehicles, this paper uses optical signals based systems mounted under the trains as an alternative approach. Deviations in the reflected optical signals are measured and interpreted on the potential cracks and appropriate alerts are generated. The system can be mounted under the train, is easy to maintain and to manage.