Development and implementation of EcoDecibel: A low-cost and IoT-based device for noise measurement

Background: Noise pollution is a growing concern for public health and the environment. Traditional noise monitoring methods often have limitations due to short-term measurements and high costs. Objective: This study aims to develop and validate EcoDecibel, a low-cost, IoT-based sensor system for continuous noise monitoring, addressing the gaps in existing noise measurement technologies. Methods: EcoDecibel was compared with Class 1 and Class 2 sound level meters in various conditions. The system was deployed across three environmental sites in Sanzhi District, Taiwan, for one week. Time-series prediction and forecasting models (SARIMA, Prophet, LSTM) were applied to the noise data to predict and forecast noise levels. Results: EcoDecibel demonstrated strong correlation, yielding R² values of 0.948 and 0.983 in comparison with Class 1 and Class 2 sound level meters and was able to monitor and forecast daily noise patterns effectively. The system performed well across different environments and was capable of continuous real-time monitoring. Conclusions: EcoDecibel presents a cost-effective and reliable solution for long-term environmental noise monitoring. The system is suitable for use in epidemiological studies investigating the relationship between noise exposure and public health outcomes.