Monitoring Noise Pollution Using the Urban Civics Middleware

Abstract

Noise pollution is a significant problem in cities due to its various effects on health, but the modeling of noise data and the generation of accurate noise pollution maps suffer from the high cost and restricted scale of sensing performed using static municipal sensors. In this paper, we present our approach for augmenting municipally sensed data using participatory sensing-based information collected from smart phones. We make use of a data assimilation method to generate more accurate noise maps that combine simulated and measured noise levels. Our solution customizes the Urban Civics middleware for noise-specific application. Urban Civics combines middleware solutions for urban-scale sensing and crowd-sourcing, and data assimilation techniques, which are the main focus of this paper, to generate, collect, and process the big data involved in this process in a scalable manner. Our experiments demonstrate the improvements in quality enabled by this technique vis-à-vis the noise map usually generated with simulation along with observational data from municipal static sensing alone or mobile sensing alone.