This study investigates the impact of automobile air conditioning systems on fuel consumption and gas emissions in hybrid and internal combustion engine vehicles using machine learning algorithms. Dynamic and static data were collected via an OBDII sensor from the vehicles examined, complemented by weather data sourced from the OpenWeatherMap API to capture external factors such as temperature and humidity during each trip. Machine learning models, including Linear Regression, XGBoost, and a Stacking ensemble model, were trained to predict fuel consumption rates over time, supplemented by an XGBoost Classifier to ascertain whether the air conditioning was in use during trips. The regressor models achieved low error scores after training and modifications, while the classifier achieved an accuracy of 99%. The findings from this research may inform future studies aimed at optimizing vehicle performance under varying road, driving, and weather conditions while minimizing energy consumption and harmful emissions.

Significant energy wastage and increased CO2 emissions from automobile air conditioning systems highlight the critical need for enhanced energy prediction and management. This study investigates the potential of machine learning algorithms for energy prediction in A/C systems, focusing on the Toyota Prius Hybrid model. By integrating vehicle data with external weather information, we developed a predictive model that accurately forecasts energy usage and identifies key factors influencing efficiency. These findings could facilitate the creation of intelligent control systems for automotive air conditioning, leading to enhanced energy savings and reduced emissions.