Logo image
Back
IoT-Driven Environmental Pollution Monitoring with a Deep Attentional Hybrid Transformer Model
Conference proceeding

IoT-Driven Environmental Pollution Monitoring with a Deep Attentional Hybrid Transformer Model

Shakir Syed, Rama Chandra Rao Nampalli, Pavithra A, Manoj Nikam, Thulasirajan Krishnan and Anvesh Perada
2025 International Conference on Emerging Systems and Intelligent Computing (ESIC), pp 356-361
08 Feb 2025
DOI: https://doi.org/10.1109/ESIC64052.2025.10962733

Abstract

Accuracy Artificial intelligence Atmospheric modeling CNN-Transformer Encoder Compounds Data models Environmental Pollution Feature extraction Training Transformers Urban areas Air Quality Water Pollution
As urbanization and globalization increase, air quality in many parts of the world is deteriorating and need quick action. Many cities have unacceptable levels of particulate matter and gaseous pollution, exceeding government and WHO guidelines. Continuous air pollution exposure increases the prevalence and mortality of respiratory disorders like asthma and COPD. Data preparation, feature extraction, and model training are the study's three steps. A neighbor-based feature scaling method prepares data for processing. Feature selection removes superfluous data before model training. The sophisticated architecture of the CTransNet model speeds up training. This strategy outperforms others, notably for attention-based models and CNNs. The model's 93.56% accuracy beats stateof-the-art alternatives. These findings demonstrate that the proposed technique can accurately predict air quality, which could enhance health by warning of imminent pollution.

Metrics

1 Record Views

Details

Logo image