An analysis of air quality in Pristina city center and its relation to traffic-derived air pollution emission

This study presents a comprehensive analysis of vehicle emissions in Pristina, Kosovo, including data collection, processing, and modelling. In the first stage, second-by￾second speed and position data were collected from five different vehicles and nine drivers using OBD devices. The collected data was cleaned and erroneous data was removed. Then, micro-trips were identified using the developed three-point consecutive algorithm and classified using the k-means clustering method. A driver cycle of 1780 seconds was created from these trips. The analysis results showed that Pristina drivers tend to drive more aggressively compared to European standard driver cycles. It was also found that average speeds at roundabouts are higher than other intersection types. The MOVES model and Monte-Carlo simulation were used to estimate emissions for 55498 registered vehicles in Pristina city based on the year 2020. According to the model based on the driver cycle, average per-vehicle emissions were found to be 48 g for CO, 1.22 g for HC, and 2.32 g for NOx. In the MOVES model integrated with SUMO software, these values were estimated as 43 g, 0.81 g, and 1.84 g, respectively. The HBEFA4 model was found to underestimate CO and HC values considerably. In conclusion, it was found that despite the increasing traffic, emission levels remained the same due to advances in engine technologies.