On the estimation of pollutants in road tunnels based on traffic data

This paper focuses on an important topic estimation and prediction of pollutants in road tunnels as an input for ventilation control systems. It demonstrates that using estimated pollutants level computed from traffic data is not only possible, but even introduces diverse advantages. Next to the ability to predict future values of the pollutants level, lower investment costs (installation of rather costly pollution sensors is not needed any more) and lower maintenance costs (minimizing costly turning on and off of the ventilation system) must be mentioned. Two different methods based on regression models and decision trees are presented and their results compared in this paper. The developed models were evaluated on data from tunnel Mrazovka located on the Prague city ring. The results clearly confirm the suitability of the proposed solution for given task.