A study of the performance of a generalized exceedance algorithm for detecting noise events caused by road traffic

A key question for road traffic noise management is whether prediction of human response to road traffic noise could be improved by accounting for noise events instead of, or in addition to, energy equivalent or percentile measures of noise exposure. However, there is a critical prior question: how should noise events in road traffic be measured? Even at moderate traffic flow rates, detecting and counting noise events caused by road traffic is not a trivial exercise, and as yet there is no generally accepted noise event detection algorithm. This paper investigates the performance of a generalized exceedance algorithm for detecting noise events, constructed on the basis of the literature on noise events caused by road traffic. For this purpose, a microscopic traffic simulation model, coupled to an emission model that accounts for distributions of sound power levels of individual vehicles, is used to simulate one-hour time histories of the noise level in the proximity of a roadway, for an exhaustive set of traffic flow/composition and propagation distance conditions in unshielded locations. The validity and reliability of the number of noise events detected by the generalized algorithm in these one-hour time histories is then evaluated for a range of algorithm parameter sets. By discarding parameter sets that do not result in an algorithm that returns valid or reliable counts, and by examining redundancy in the remaining ones, a small number of representative parameter sets is identified, which may prove useful in the construction of event-based indicators supplementary to energy-equivalent measures of road traffic noise.