REAL-WORLD ELECTRIC VEHICLE ENERGY CONSUMPTION PREDICTION USING MACHINE LEARNING ALGORITHMS

Abstract

The rapid adoption of electric vehicles (EVs) has increased the need for accurate prediction of energy consumption to improve battery management, driving efficiency, and route planning. Electric vehicle energy usage is influenced by multiple factors such as driving behavior, vehicle characteristics, road conditions, traffic patterns, and environmental variables. Traditional estimation methods often fail to capture these complex relationships, leading to inaccurate predictions. This study proposes a machine learning–based approach for predicting real-world electric vehicle energy consumption using field data collected from EV sensors and telematics systems. The system analyzes various input features including vehicle speed, acceleration, battery state of charge, temperature, road gradient, and driving patterns. Machine learning algorithms such as Random Forest, Support Vector Machine, and Gradient Boosting are applied to model the nonlinear relationships between these factors and energy consumption. The proposed framework preprocesses the collected data, extracts relevant features, and trains predictive models to estimate energy usage under different driving conditions. The trained models are evaluated using performance metrics such as Mean Absolute Error (MAE), Root Mean Square Error (RMSE), and prediction accuracy. Experimental results demonstrate that machine learning algorithms can significantly improve the accuracy of EV energy consumption prediction compared to traditional methods. The developed system can support intelligent energy management, optimal route planning, and improved battery utilization in electric vehicles. This research contributes to the development of smarter and more efficient EV transportation systems.