Development of Warning and Predicting System for Quality of Air in Smart Cities Using RNN

In this paper, we propose to build an intelligent air quality warning and prediction system for smart cities. The system will automatically provide information of temperature, humidity, dust concentration, and UV rays in defined areas. The system includes two parts, namely sensor and communication blocks. Sensor block will collect information and transfer to the server that performs to forecast and display the parameters (temperature, humidity, dust concentration, and UV rays). In communication block, the data is transmitted by SIM module to website to provide information and warn to the user if finding any problems. In our system, we use recurrent neural network (RNN) to forecast the change of the parameters. The results show that the proposal system improves the accuracy above 90% comparing to real system.