Accuracy impact analysis for speed-based dynamic updates of regional road-traffic noise emissions

Abstract

Dynamic updates of regional road traffic noise emissions are essential for generating accurate dynamic noise maps. However, existing research has not thoroughly analyzed the accuracy of key components: road-segment classification and update methods. This study addresses this gap by using high-temporal-resolution traffic speed and noise data. Specifically, we constructed 34 types of classification features from speed sequence data, with each type having 4 forms, to categorize unmonitored road segments into categories of monitored road segments based on the availability of noise-related data. We applied the direct update method and difference-based update method families to periodically update noise emissions. We employed the leave-one-out cross-validation method to verify the accuracy of various combinations of features and update methods, and analyzed their error sources. Based on the findings, we provide recommendations for dynamically updating regional road traffic noise emissions based on speed data, offering valuable insights for policymakers in environmental management.