Development of optimized algorithms for the classification of networks of road stretches into homogeneous clusters in urban areas

Noise maps are considered as a powerful tool to determine the population exposure to environmental noise. A statistical approach to real-time noise mapping will be developed in DYNAMAP (Dynamic Acoustic Mapping), a co-founded project in the framework of LIFE 2013 program. The main preliminary action of the project is to define a statistically-based method to optimize the choice and the number of monitoring sites, which will provide the information to update the dynamic mapping process. In this work, preliminary results referring to a sample of roads of the city of Milan are presented. The sample database is made of 24 hour continuous acoustic monitoring of the hourly equivalent levels LAeqh in different sites, corresponding to 8 road functional classifications (from A to F and sub-classes). Once normalized, such trend profile provides a tool to group roads by their vehicular dynamics. Acoustic trend profiles will be also studied on a shorter time basis, with the aim of identifying road clusters that allow an updating of the map with an higher time frequency. Linking a non-acoustical parameter (hourly traffic flow) to the elements in each cluster represents the key-issue which allows each road segment of the urban network to be univocally assigned to the obtained clusters.