A fuzzy expert system capable of computing LA,max for the Tehran-Karaj commuter train

This paper describes the methodology through which a fuzzy expert system capable of computing maximum A-weighed noise level (L_A,max) for the Tehran-Karaj commuter train is developed, tested, tuned and validated on the basis of the 50 measurements from five locations at distances of 25 m, 35 m, 45 m, 55 m, and 65 m from the centre of the track and at a height of 1.5 m. The expert system is initially developed, tested and tuned through the judgements made on the premise of the samples taken at the distances of 25, 45 and 65 m. The conducted hypothesis test, the non-parametric two-related samples Wilcoxon test, for this stage indicates satisfactory results. Eventually the developed expert system is validated for the measurements related to the distances of 35 and 55 m through carrying out the non-parametric two-independent samples Kolmogorov-Smirnov test.