Smartphones can be used to collect measurements of noise levels and improve urban noise maps. The mobile observations and traditional noise maps (as computed by noise mapping models) can be merged using data assimilation methods. These methods can process high-dimensional data and often compute unbiased estimators with minimum error variance.

Below is an example of a simulated noise map corrected by mobile observations collected with smartphones. There was a special event that day which drastically limited road traffic, hence the lower noise levels.

This work was developed and included as part of the technology of the startup Ambiciti. Other technologies I developed were used by the startup for its pollution API or for its mobile application (Android & iOS, see below), such as statistical modeling of air pollution at street resolution and optimal routing to minimize exposure to (noise or air) pollution.

The assimilation work was also emphasized by the Journal of the Acoustical Society of America, and selected for its cover (below).