Locate to innovate: the link between traffic density and pollution revealed

Our cities are getting more polluted, and there is a serious cost to our health and wellbeing. Ambient air pollution causes about 4.2 million premature deaths every year worldwide, according to some estimates. Stroke, heart disease and lung cancer are just some of the ways people die because of poor air quality.

“For cities and governments, air quality monitoring can be used to identify areas with high levels of air pollution, and optimize urban circulation policies to improve air quality," explained Xuening Qin, a researcher at Ghent University and imec in Belgium.

Citizens can also use this information to make decisions such as where to jog, whether to open or close windows or when to stay inside to avoid an asthma attack.

But traditional models of measuring air quality are far from perfect. Stationary monitoring stations only provide average levels over large areas. They are highly accurate but expensive, and — most importantly — do not show how pollution levels vary within that area.

Local sources such as industrial emissions or emissions from vehicles can cause pollution hotspots hidden in the average pollution number.

Other factors, such as the street type, wind speed and direction, and temperature, also play a role. So-called urban canyons, when roads are surrounded by tall buildings and can lock in air pollution when the wind is blowing in the wrong direction.

The challenge

What Xuening and her team wanted to do was create a low-cost and more accurate model to estimate local air pollution by linking the estimates to the local context. They used mobile sensors attached to postal vehicles — but an added ingredient was to make all the difference.

“Because the air quality measurements created by the mobile sensors are sparse, but traffic data is abundantly available, we could use that to greatly improve the resolution and accuracy of the pollution estimates without magnifying costs," she told HERE360.

They used traffic data donated by HERE to create a new model for mapping air quality.

And the results? After testing the air in Antwerp, Belgium, using this model in a three-year imec research, the team was able to produce a fine-grained air quality map. It shows clearly how pollution levels change over close distances and short periods of time.

“The relation of traffic density with air pollution is very real," she said. “We think real-time traffic data is one of the most important contextual features for our model."

When compared to traditional fixed monitoring systems for three sites — a roadway, a highway and a street canyon — the new model was found to be more accurate in every case.

What's next?

The map Xuening's team has produced shows not only where the air pollution hotspots are, but also the factors that might have led to this. It can be used to implement more environmentally-friendly traffic policies, such as preventing cars from driving around school streets at times children are walking to and from school.

And it has commercial applications as well. As long as consumers are willing to pay, the air quality mapping model can be used by individuals deciding where to live, exercise and more. It can even be used to predict what the effects of changes to the urban environment might be on air quality.

That has to be good news for the millions of people worldwide affected by air pollution and the problems it causes.