Air Quality

Cities would like to use networks of low-cost sensors for air quality monitoring.

Such sensor networks can provide very detailed information about pollution levels across a city. However, the sensors are not as accurate as more expensive equipment. 

To address this, we will develop new ways to combine and process data from many sensors to get better overall results. Methods will be created to automatically detect sensor errors and aging over time. Advanced statistics will be used to quantify the reliability of measurements.

These methods will be tested on a large network in the Paris region consisting of 600 units. 100 are at fixed locations and 500 are mounted on postal vehicles, mapping pollution street-by-street. A wide range of pollutants are monitored: different sizes of particulate matter, nitrogen dioxide, carbon monoxide and ammonia.

The project partners will provide the network operators with uncertainty-aware city-wide particle pollution maps. The goal is to significantly reduce measurement uncertainty for all monitored pollutants.

If successful, the project will enable city-wide air quality monitoring at low cost, with measurements reliable enough for decision-making. This could help cities around the world better understand and address air pollution.