Study of Weight in Motion Sensor for Railroad Crossing Warning System Using Artificial Neural Network

Abstract

Abstract IThe risk of accidents in the railway, especially on unmanned railroad crossing, are still high. A train arrival on railroad crossing warning system is indispensable. The warning system will detect train arrival on the railroad crossing and issue a warning to the crossers. A sensor based on optical fiber Weight in Motion (WIM) is proposed. This paper presents a design of a railroad crossing warning system based on WIM sensor. The sensor placed under the railroad tracks and detects vibrations footprints transmitted along those rails. Artificial Neural Network (ANN) is used to process the signal pattern of the WIM system output. By using ANN, the vibration footprint pattern was classified and used as on-sensor train detection. As the train detected, the arrival time of the train on the raiload crossing was estimated. The result is a warning system which can detect train arrival on the railroad crossing and issue a warning signal.