غفران عصام دريول
Air pollution is a leading cause of health concerns and climate change, one of humanity's most dangerous problems. This problem has been exacerbated by an overabundance of automobiles, industrial output pollution, transportation fuel consumption, and energy generation. Protecting people from air pollution damage is an important issue in smart cities and to accomplish this task by designing an integrated system for monitoring, forecasting and displaying air pollution data. The system is further divided into three subsystems: a monitoring, prediction and data virtualization. pollution levels are observed in three cities in Baghdad using different types of sensors connected with ESP32 to detect Particulate Matter (PM2.5 and PM10), Nitrogen Oxides (NOx), Carbon monoxide (CO) to monitor outdoor air quality. Observed results are monitored by ThingSpeak, which is an opensource IoT platform. For prediction, a deep learning model was proposed using the LSTM algorithm and the GA optimization algorithm to find the best LSTM hyperparameters. The results showed that the GA optimization algorithm significantly improved the LSTM prediction by confirmation of the prediction results of the proposed model by comparing it with other prediction models. Two types of evaluation metrics RMSE and MAE were used, the results were respectively 9.5820 and 19.164. Finally, a website was built to classify air pollution from good to Hazardous for inhalation to display the air quality index data. This thesis contributes to finding a solution to the problem of air pollution in cities.