Classification and clustering-based AI/ML applications in driving behavior analysis and vehicle energy consumption: A systematic literature review

The transportation sector is a significant contributor to global energy consumption and greenhouse gas emissions. Artificial intelligence (AI) offers strong potential to optimize vehicle energy efficiency by analyzing complex driving behaviors. However, the increasing complexity of AI models has created a need for explainable AI (XAI) to ensure transparency, trust, and adoption. This systematic review aims to synthesize evidence on classification and clustering-based AI/machine learning (ML) studies used to analyze driving behavior and its relationship with vehicle energy or fuel consumption. The review includes studies that use classification and/or clustering as a main AI/ML component to examine the relationship between driving behavior, whether human or automated, and vehicle energy or fuel consumption. Out of 207 initial records, 53 studies met the inclusion criteria. Common AI techniques include supervised learning (e.g., support vector machines) and unsupervised learning (e.g., k-means and self-organizing maps). Key driving features affecting energy consumption include speed, acceleration and deceleration patterns, and driving volatility. Aggressive driving styles are consistently associated with higher energy consumption. XAI methods, such as Shapley additive explanations, are used to interpret AI model predictions and identify influential factors. However, their integration into real-time driver feedback systems remains limited. AI-driven interventions, including eco-driving assistance systems and adaptive energy management strategies, show measurable improvements in energy efficiency. Future research should focus on real-world validation, standardized evaluation metrics, driver-centered XAI applications, and the integration of diverse data sources to improve both energy efficiency and safety.