IoT Based Railway Track Faults Detection and Localization Using Acoustic Analysis

Rail is one of the most energy efficient and economical modes transportation amongst many. Regular inspection railway track health essential for ensuring robust secure train operations. Investigations delayed discovery pose a serious risk to safe functioning rail transportation. The traditional method manually examining using cart both inefficient susceptible mistakes bias. It imperative automate in order avert catastrophes save countless lives particularly zones where accidents are numerous. purpose this research develop an Internet Things (IoT)-based autonomous fault detection system enhance existing address aforementioned issues. In addition data collection on Pakistani lines, work contributes significantly identification classification based acoustic analysis, as well localization. Due their frequency occurrences, six types track’s faults were first targeted: wheel burnt, loose nuts bolts, crash sleeper, creep, low joint, point crossing. Support vector machines, logistic regression, random forest, extra tree classifier, decision multilayer perceptron ensemble with hard soft voting among machine learning methods used. results indicate that can successfully assist discriminating defects localizing real time. show MLP achieved best results, accuracy 98.4 percent.