A data-driven approach to fuel consumption analysis of logistics vehicles in Beijing

As the world increasingly prioritizes green, low-carbon, and sustainable development, reducing transportation-related pollution has become a focal point of research for many scholars. This study focuses on examining the fuel consumption of logistics vehicles operating along the Sixth Ring Road in Beijing. To begin, we categorize the driving behaviour into segments based on changes in vehicle speed. These segments are then clustered using the K-means algorithm, which identifies three distinct driving states. Next, we build a fuel consumption prediction model using the Long Short-Term Memory (LSTM) algorithm, leveraging historical fuel consumption data associated with different road types corresponding to each state. After conducting our analysis, we find that logistics vehicles exhibit higher fuel consumption per 100 km, as well as greater fuel consumption fluctuations per second, particularly on branch roads characterized by complex road conditions. The insights derived from this study contribute to improving energy efficiency management for medium and heavy trucks.