An Intelligent Humidity and Temperature Monitoring System using Long Short-Term Memory Neural Network and Raspberry Pi

Monitoring systems often use humidity and temperature readings to predict weather conditions in certain areas. However, these climatic factors also correlate with agricultural impacts and public health, which may affect cities' functioning. An intelligent humidity and temperature monitoring system is therefore proposed that can be easily deployed in remote and targeted areas to help stakeholders devise appropriate countermeasures to mitigate climatic impacts. An ESP32 microcontroller integrated with DHT11 and BMP280 sensors was used as a standard component of the system due to their availability. These components can record and store data on a NoSQL database to allow remote monitoring and sensing. The Raspberry Pi integrated an LSTM neural network to predict the maximum temperature and humidity values based on the previous three days' humidity, maximum temperature, and minimum temperature readings. Results have shown that the proposed system could closely predict the values of maximum temperature and humidity. Future research may include adding more sensors to create a wireless sensor network for improved data gathering and monitoring procedures.