Noise mapping based on participative measurements with a smartphone

Abstract

International audienceBecause noise is a major pollution leading to non-negligible socio-economical impacts, many national regulations aim at reducing the population noise exposure. Within the context of the European directive 2002/49/EC, a special attention is paid to the evaluation of the existing noise environment. Nowadays, this assessment is addressed based on simulated noise maps, which however present some limitations due to the simplification of noise generation and propagation phenomena. Smartphone participative measurements are alternatively being developed, offering the high temporal and spatial granularities recommended by the EU directive. However, the existing approaches often lack a quantification of the produced noise maps accuracy, and are rarely user-oriented. In this context, within the framework of the EU project ENERGIC-OD, a Spatial Data Infrastructure (SDI “OnoMap”) has been developed to manage smartphones measurements using a dedicated Android application (“NoiseCapture”) and to produce relevant noise maps. In the present communication, this infrastructure is detailed, with a specific attention to the following major key points: data management and qualification from its production to its dissemination, use of standards for data interoperability, optimization of noise measurements, production of 24h noise maps based on measurements distributed over a day, integration of a confidence index on the produced data

Road traffic sound level estimation from realistic urban sound mixtures by Non-negative Matrix Factorization

“…In response to the growing demand from urban dwellers for a better environment, noise mapping has been recom- Noise measurements are thus increasingly used in addition to simulation to describe urban noise environments [5][6][7]. Several measurement set-ups have been proposed in the last years, including mobile measurements with high quality microphones [8,9], participative sensing through dedicated smartphone applications [10,11], or the development of fixed-sensor networks. In this latter case, the sensor networks can be based either on high-quality sensors as in [12,13], or low-cost sensors as in the DYNAMAP project [14] or the CENSE project [15].…”