Evaluation and Prediction of Climatic parameters based on IoT system using Machine learning and Neural network algorithms

Abstract

This study presents an innovative method for integrating artificial intelligence/machine learning models with an Internet of Things (IoT) system for evaluating, validating and predicting climatic parameters. The prototype IoT system described herein is designed to remotely control demo-satellite power-enabled technologies while simultaneously capturing real-time environmental parameters such as air temperature, humidity, and atmospheric carbon dioxide (CO2) levels. These data are then stored in a dedicated server database. They are capable of interfacing with the Blynk IoT application for seamless remote-control access of demo-satellite and monitoring environmental parameters; However, the existing system suffers from low resolution and accuracy in data collection. To address this limitation, we propose high-resolution ground data for the implementation of AI/ML models to improve the accuracy of atmospheric air temperature and CO2 level predictions each algorithm's performance is assessed using metrics like Root Mean Squared Error (RMSE), Mean Squared Error (MSE), and Percentage Error (PE) and Accuracy percentage. Additionally, future predictions for time horizons of 1 month, 6 months, and 1 year are displayed graphically in the system's Graphical User Interface (GUI). This paper presents a comprehensive overview of the proposed system architecture, implementation details, experimental methodology, results analysis and predictions.