Evaluation of leisure noise in urban environments

an approach based on low-cost sound monitoring systems and artificial intelligence

Abstract

With the growing concern about noise pollution and its effects on health and well-being, the development of effective urban sound planning is becoming highly necessary. However, the soundscape of the urban settlements is a complex composition of different sound sources and should be faced with a multidisciplinary approach. The urban sound planning concept addresses not only the reduction and time of exposure to noise, but also the subjective and qualitative nature of sound, which enables the design of favorable acoustic environments. In this context, one sound source that stands out is the leisure noise. Being composed by diverse sources, it is hard to be characterized, and standardized prediction or measuring methods are still under development. However, it represents essentially the liveliness of the city and the important actions of relaxation and socialization of their citizens. Different international research study how sound monitoring stations integrated with the latest technologies could improve the collection of quantitative data regarding recreational noise, other studies have developed questionnaires to understand its subjective aspects. Furthermore, in São Paulo, a low-cost sound sensor was created with the aim of facilitating urban noise monitoring. An overview of its development process will be presented in this article, as well as the evolution of machine learning for urban sound source recognition.