Innovative procedure for measurement uncertainty evaluation of environmental noise accounting for sound pressure variability

This study aims to demonstrate the importance of uncertainty evaluation in the measurement of environmental noise in the context of Italian legislation on noise pollution. Attention is focused on the variability of the measurand as a source of uncertainty and a procedure for the evaluation of uncertainty for environmental noise measurement is proposed. First drawing on several real noise datasets in order to determine suitable measurement time intervals for the estimation of the environmental noise, a data-driven sampling strategy is proposed, which takes into account the observed variability associated with measured sound pressure levels. Outliers are eliminated from the actual noise measurements using an outlier detection algorithm based on K-neighbors distance. As the third step, the contribution of measurand variability on measurement uncertainty is determined by using the normal bootstrap method. Experimental results exploring the adoption of the proposed method drawing upon real data from environmental noise using acquisition campaigns confirm the reliability of the proposal. It is shown to be very promising with regard to the prediction of expected values and uncertainty of traffic noise when a reduced dataset is considered. [edited by author]