Urban noise pollution is one of the major environmental challenges affecting public health, comfort, and the livability of modern cities. Traditional noise monitoring methods are limited due to their high cost and low spatial coverage. This research introduces an IoT-based intelligent system for real-time urban noise monitoring and prediction using Artificial Intelligence (AI). The proposed architecture consists of a distributed network of low-cost IoT sensor nodes, built on Raspberry Pi platforms, deployed across urban environments to capture high-quality acoustic data. The collected data is transmitted to a cloud-based platform, where a deep learning model — specifically a Convolutional Neural Network (CNN) — is used for environmental sound classification and short-term noise level prediction. The system not only identifies major noise sources but also predicts future noise intensity trends and automatically generates alerts when noise thresholds are exceeded. Simulation results demonstrate high accuracy in sound classification and effective short-term prediction performance. This work contributes a scalable and cost-effective solution for dynamic noise mapping, supporting smart city initiatives and improving urban environmental quality.

Urban Noise Pollution, Internet of Things (IoT), Artificial Intelligence (AI), Deep Learning, Convolutional Neural Network (CNN), Smart Cities, Environmental Sound Classification