PROCEDURE TO DEVELOP AN ARTIFICIAL NEURAL NETWORK MODEL THAT PREDICTS ANNOYANCE TO TIME-VARYING NOISES

摘要

The development of a metric to predict the human reaction to noises has been of interest to noisecontrol engineers for many years. However, the “quality” of the sound that machines generate hasreceived significant attention in recent years. The goal of the research presented in this paper was tobuild and train an artificial neural network (ANN) t o g enerate a robust model that predicts humanannoyance to noise stimuli where the inputs to the model are obtained from analysis of the noises.It is anticipated that this neural network will be used by engineers to identify the components ofmachinery noise that contribute most to annoyance. With this knowledge, engineers would thenbe able to modify the design of the machinery in order to minimize noise annoyance.