Fuel efficiency and driving behavior analysis play an important role in improving modern transportation systems. Vehicles equipped with an Engine Control Unit (ECU) generate continuous operational data that can be used to understand vehicle performance and driver behavior. This study proposes a framework based on Machine Learning to analyze real-time data collected from the Engine Control Unit for predicting fuel consumption and identifying different driving profiles. Key vehicle parameters such as engine speed, throttle position, fuel injection rate, and vehicle speed are used to train predictive models. In addition, clustering techniques are applied to categorize driving patterns into eco, normal, and aggressive driving behaviors. The developed models demonstrate high prediction accuracy and provide meaningful insights into driving patterns. The proposed system can support intelligent transportation applications by helping drivers and fleet managers monitor fuel usage, improve driving efficiency, and promote more sustainable driving practices.
Sangannagari et al. (Sun,) studied this question.