Combating air pollution: cleaner air for Brixton

“Two degrees Celsius and light showers.” Confronted with a weather forecast like this, most people instinctively know what to do: dress warmly and don’t forget to take an umbrella. The sophisticated sensor system developed by Dr. Maria Martínez Prada and her team also generates forecasts. But in this case, their charts and graphs reveal more than just the overall weather situation. They also provide a battle plan against air pollution, and can contribute to a better quality of life in urban areas. While there’s generally no changing the weather, environmental pollution is certainly something that can be mitigated.

Take London: it’s a great city with a lot of flair — but it also has a major problem with nitrogen dioxide (NO₂) emissions. On Brixton Road, for example, a mainly two-lane street south of the River Thames, NO₂ levels regularly exceed the EU-wide limit of 40 micrograms per cubic meter. A partnership between Transport for London (TfL) and the Bosch Connectory, a development center for solutions relating to the internet of things, is seeking to change this with a service aimed at optimizing traffic flows to reduce air pollution. A key role here is played by an innovation that goes by the simple name of “air quality solution.”

Once in the cloud, this data is correlated with other data, such as traffic volume. As their name suggests, however, the air quality monitoring boxes only measure ambient air quality — in other words, the concentration of pollutants such as nitrogen oxides that are present in the air in the box’s vicinity. They don’t measure what individual cars emit. Particulates emitted by vehicles, for example, are not deposited equally everywhere.

Some adhere to the ground, while others are blown away by the wind or washed away by rain. The monitoring boxes thus do not record total vehicle emissions, but rather the proportion of them that remain in the air at heavily impacted locations.

Since the boxes detect all airborne emissions regardless of where they originated, the air pollution cannot be traced back to specific sources. To be able to identify these sources nonetheless, further parameters are factored into the calculations performed in the cloud, such as data about traffic flow and congestion. This data is provided by Here Technologies, a geodata company in which Bosch holds a stake. Here gathers the data from sources such as onboard vehicle navigation systems, and supplies the data in anonymized form. In addition, TfL provides information on traffic-light phases and vehicle types. Bosch has a wealth of expertise about the powertrain technologies of diverse vehicles, and can use this information to model emissions levels. In addition, parameters such as satellite data on meteorological conditions and information from the municipal authorities on the topology of buildings are collected. In a dispersion simulation, the combined data are translated into charts and graphs.

To forecast pollution levels, an algorithm based on artificial intelligence is used. The algorithm was generated by means of machine learning and takes data collected by individual monitoring stations over the last two years as its starting point. On this basis, the algorithm can calculate the impact that specific weather conditions, rush-hour traffic, different days of the week, and major events can have on air quality levels in any given area. By combining empirical data and the measurements from the air quality measurement boxes, the AI can forecast what effect traffic-related emissions will have on air quality in certain weather conditions. This knowledge can be used to intervene in traffic management well in advance.

Pointing to numbers displayed on a big screen, Martínez Prada explains how this works: “This data tells us precisely how much traffic is contributing to local air pollution. We can do this by comparing emissions during a weekday rush hour with those on the weekend, for example, or by analyzing the impact of bad weather on particulate levels. The data will even tell us what impact a light east wind will have on air quality.”

These precise forecasts are the product of a wealth of multidisciplinary expertise at Bosch: the company possesses not only vast knowledge in sensor technology but also decades of experience in dealing with vehicle emissions. Moreover, it knows how to use AI to process data. But how will all this benefit the people on Brixton Road? This is where TfL plays a key role. A service created on the basis of the air quality solution can be used to test whether, and to what extent, measures such as rerouting traffic and modifying traffic-light phases can help improve traffic flow, and thus air quality.