Predictive Weather Modelling Using Unmanned Aerial Systems and Machine Learning

One amazing benefit of science and technology is weather forecasting, which is essential to daily life. Predicting the weather is a challenging task because the environment is a continuous, multifaceted process. Various data-driven techniques are necessary for weather information collection and processing. These techniques can combine the relatively simple view of the sky with complex numerical models to produce the predictions. This paper presents different algorithms, which applied the concepts of Decision tree, KNN, Gradient Boost, Random Forest regressor, in order to forecast temperature variations, including high and low values as characteristics in a linear combination. Unmanned Aerial Vehicles (UAV) can be used for prediction for weather in a remote area. The aim of this work is to develop a UAV that can stay in “flight mode” at a fixed altitude for a certain amount of time and measure the weather-related parameters at that geographical location and can send the measured data to the ground station. The primary parameters focused are the highest and lowest temperatures, average humidity levels, and rainfall level. A light sensor is used to determine the amount of daylight, and a sensor for measuring air pressure is also employed. Weather forecast can be predicted by machine learning using parameters such as altitude, humidity, temperature, rain, pressure.