Inertial Events Identification Using Neural Networks in Railway Environments

Global positioning system (GPS) resolves railway vehicles location most of the operation time. But what happens when those vehicles are near to critical points, as rail track changes, where more precision is needed? This paper presents a possible solution using inertial sensors and neural networks to find out patterns of railway changes among the inertial sensors' signal. The main objective of this investigation is to determine whether is possible or not to design a neural network capable of identifying or classifying the inertial patterns resulting from the trainpsilas railway changes.