Railway Antiskating Systems Controlled by Neural Networks

In order to maintain a good level of safety, the electronic components have to check the functioning of the mechanical parts and if it is the case some suitable elements can be controlled so that the standard safety level is aleways guaranteed. The aim of this paper is the simulation of a railway anti-skating system by using a Neural Network as checker of the whole system. The Neural Network takes care of the critical stage of the process that is when all wheels are simultaneously skating. In this case in fact it is impossible to determine the train reference speed within an acceptable error limit. For the resolution of this problem a Neural Network based on a new learning strategy, called M.A.I.A, has been employed. This kind of Neural Network is generally quick and efficient.