Origin–destination specific traffic emissions and data-driven NO2 pollution-optimal routing in urban environments

Urban air pollution from traffic poses serious public health risks. Pollution exposure can be minimised through traffic routing systems; these currently rely on detailed local environmental information, which is often difficult to collect or generalise within and across cities. Here, we introduce a new data-driven approach for ready application to different urban road networks by directly relating NO2 to traffic density in a time-dependent and weather-corrected manner. We demonstrate this application by comparing pollution-optimal routings, using our novel direct NO2/density approach, to the conventional traffic assignment minimising user travel time, in a case study of Sheffield, UK. There, we find user-optimal traffic flows result in 21% higher total NO2 concentrations than pollution-optimal routings, while saving only 9% in total travel time: an average of 0.3 min per road. Our generalisable framework offers a practical alternative to current emissions-based models for air-quality-aware traffic control and environmental zone planning.