Globally, vehicular pollution is one of the greatest concerns in urban areas. Several studies on air pollution have been conducted using deterministic, statistical, and soft computing methods. However, there has been little research on how soft-computing methods like Artificial Neural Networks (ANN) can help us comprehend vehicular pollution’s non-linear and highly complex dispersion. This study uses an ANN-based vehicular pollution model to investigate the effect of vehicular traffic on PM2.5 concentrations in built-up and open terrain-surrounding environments. Five distinct pollution models were developed for two locations in Delhi, considering PM2.5 pollutants, meteorological variables, traffic flow, and traffic composition into account. The results concluded that under open terrain conditions, the significance of the traffic variable in its association with PM2.5 is almost half the significance observed under built-up conditions. Also, in terms of PM2.5 reductions, the maximum reduction observed at Location-1 (built-up environment), and Location-2 (open terrain environment) is 1.85 and 2.44 times the percent reduction in traffic during peak hours, respectively. The study’s findings have significant ramifications for the current practices of ignoring the contribution of traffic and the built environment to pollution and adopting measures like an odd-even rule and high fuel and parking prices to combat pollution.
Read full abstract