Regresar

Sound noise levels calibration with smart-phones

Abstract:

Recent research suggests that there are adverse effects of loud noise on people in urban areas. This makes noise assessment an important concern for policy-makers, urban planners, and citizens in general. Despite the existence of standards for sound noise measurement, there is not a common strategy for data collection over a long period. The purpose of this study is to validate the use of smart-phones to measure noise in the context of a crowdsensing application. Validation will be performed by comparing noise measurements obtained by smart-phones to measurements obtained by using a calibrated Sound Level Meter. We have performed calibration of five models of smart-phones using android Operating System against a calibrated Libelium Waspmote Smart-Cities device with integrated microphone. Four methods of calibration were tested: Linear Regression, Interpolation Table, Quadratic Regression, and Laplace Transform. The findings show that the RMSE between phone measurements and the calibrated device is higher of 4 dB on a validation data set from a real case. An adaptive calibration method is needed.