Machine Learning Applications in Energy Consumption Forecasting and Management for Electric Vehicles: A Systematic Review
Emilia Szumska, Łukasz Pawlik, Damian Frej, Jacek Łukasz Wilk-Jakubowski
Abstract
This literature review addresses a major research gap in electromobility by providing a comprehensive synthesis of machine learning (ML) and deep learning (DL) applications for forecasting energy consumption, managing battery state of charge (SoC), and integrating electric vehicles (EVs) with charging infrastructure and smart grids, including vehicle-to-grid (V2G) systems. Despite the rapid increase in publications between 2016 and 2025, few comparative studies systematically evaluate ML/DL approaches, their effectiveness in specific applications, and their limitations under real-world conditions. To bridge this gap, this review analyzes 95 publications, covering methods from ensemble learners (e.g., Random Forest, XGBoost) to advanced hybrids (e.g., LSTM + MPC). Key influencing factors such as driving style, topography, and weather are considered. This review identifies persistent challenges, including the lack of standardized datasets, limited model generalization, and high computational demands. It also outlines research directions, such as adaptive online learning and integration with V2X technologies. By consolidating current knowledge, this review supports engineers, EV system designers, and policymakers in planning effective energy management and charging strategies, thereby contributing to the sustainable development of electromobility.