Smartphone measurements map neighborhood noise

With their fine spatial detail, simulated maps of background noise from road traffic, aircraft, railways, and the like can pinpoint noise pollution hot spots. Such maps guide policymakers toward ways to alleviate the psychological and physical disorders of excess noise. But some simulation inputs, such as traffic flow and vehicle speed, can be difficult to estimate accurately and precisely. In a new approach, Raphaël Ventura and colleagues at Inria , the French national institute for computer science and applied mathematics, used the now ubiquitous smartphone to collect more realistic noise measurements and improve the simulations’ estimate of true noise pollution. For several days, a single participant walked the streets of a Paris neighborhood for one hour in the morning and another in the evening. Using the Ambiciti mobile application, the phone’s microphone gathered noise levels every five seconds while the GPS recorded the time and location (see the figure). To enhance the simulations, the researchers applied a statistical procedure designed to minimize the difference between the smartphone measurements and the estimated noise state. Assimilating the data with the simulations reduced the error of the noise map by 50% for the two busiest streets. Moreover, the researchers took advantage of the simulations’ increased temporal resolution to produce an hourly map. They discovered that noise levels during evening rush hour were higher than the morning’s, especially on the study area’s two major streets. (R. Ventura, V. Mallet, V. Issarny, J. Acoust. Soc. Am. 144, 3, 2018.)