Scalable measurement of air pollution using COTS IoT devices

Air pollution levels have been rising at an alarming rate for the past ten years. The situation is considerably worse in developing nations, such as India. The average concentration of PM10 in Delhi has increased by over 66% between the years 2007 and 2010 and continues to increase further. Rising air pollution has been shown to have a detrimental effect on human health. The first line of action is to sensitize people about the problem by informing them about the quality of air that they are breathing in their immediate vicinity. Unfortunately, India still lacks the infrastructure required to measure pollution at a granular scale. Most of the pollution monitoring stations are placed in regions of low population density, and hence, it is difficult to calculate the personal exposure to air pollution for most of the population. It is also not economically viable to add pollution monitoring devices at such a scale in a short period of time. We propose a framework to estimate air pollution for a given locality by leveraging the existing infrastructure of monitoring stations and looking at factors, such as traffic conditions and greenery. We evaluate our framework by estimating the pollution exposure for long trips undertaken by users, given the seed pollution values at a few spots. Our framework incurs a reasonable accuracy. We find that greenery has more impact on pollution than traffic conditions.