Modeling the Relationship Between Traffic Intensity and Urban Air Pollution with LSTM Networks

Surmeli N. U., Alptekin G. I.

11th Intelligent Systems Conference, IntelliSys 2025, Amsterdam, Hollanda, 28 - 29 Ağustos 2025, cilt.1567 LNNS, ss.66-83, (Tam Metin Bildiri) identifier

  • Yayın Türü: Bildiri / Tam Metin Bildiri
  • Cilt numarası: 1567 LNNS
  • Doi Numarası: 10.1007/978-3-032-00071-2_4
  • Basıldığı Şehir: Amsterdam
  • Basıldığı Ülke: Hollanda
  • Sayfa Sayıları: ss.66-83
  • Anahtar Kelimeler: Air pollution, Air quality, Intelligent transportation systems, LSTM, Smart cities
  • Galatasaray Üniversitesi Adresli: Evet

Özet

The increasing impact of traffic on urban air quality poses significant challenges to public health and environmental sustainability. This study investigates the effects of traffic on air quality in Istanbul, with a focus on PM10, CO, and NO2 concentrations. Using data collected over 26 months, including hourly average speeds and pollutant levels, a predictive Long Short-Term Memory (LSTM) model was developed to capture temporal patterns in traffic and pollution data. The model demonstrates strong predictive capabilities for hourly pollutant levels under varying traffic conditions, validated through multiple performance metrics. The results reveal that traffic density significantly influences air quality, with higher vehicle speeds correlating with lower pollutant concentrations. Conversely, slow traffic and congestion contribute to pollution accumulation in urban areas. These findings emphasize the critical role of traffic management in mitigating urban air pollution. By combining machine learning techniques with continuous monitoring, this study provides actionable insights to develop effective and sustainable strategies for improving urban air quality.