Abstract
As most of the world population live in cities with over 60% increase in population in 2030 presenting major challenges related to noise pollution that leads to health concerns, there is a demand for a smart solution that mitigates noise pollution in cities before it reaches dangerous levels. In this paper, we propose a smart city architecture through the internet of things (IoT) to mitigate noise pollution. Mitigation includes the detection and reporting of affected areas and the prediction of possibly affected areas in a city. To evaluate the proposed architecture, we designed Noise Pollution Monitoring System (NPMS) that locates the areas, the time, the level of noise and any related events in the area to provide intelligence throughout applying data mining techniques. Intelligence includes determining if a city is suffering noise pollution, frequent locations where noise pollution needs to be mitigated and the time typically noise pollution occurs in correspondence with nearby occurrences e.g. rush hour, or reported incidents (events) in the city local news agencies. We designed noise pollution detection endpoints that can be deployed at certain locations to collect information including location, time and sound volume. We deployed the endpoints at two major intersections to collect noise levels in order to detect if an area suffers noise pollution and then utilized Support Vector Machine (SVM) and Random Forest classifiers to create a model of prediction of noise pollution before it occurs. Finally, we report the validity of prediction algorithm as well as the functional system.