This study aims to detect and analyze sustainable driving styles with respect to road safety and fuel economy using real-world trip data collected via smartphone sensors. A two-level clustering approach was applied using K-means: first, trips were segmented by average speed and road type share; second, behavioral indicators such as harsh braking events, mobile phone use, acceleration variability, and fuel consumption were used. Principal Component Analysis was applied for dimensionality reduction, and the Silhouette method for optimal cluster selection. The analysis revealed distinct driving profiles across urban, rural-dominant and highway-oriented driving. In rural settings, safe and eco-driving behaviors were strongly aligned. On highways, fuel efficiency sometimes coincided with riskier behaviors such as distraction. In urban contexts, however, some less fuel-efficient drivers exhibited relatively safer behavior, suggesting a trade-off. These findings underscore the context-dependent nature of sustainable driving and highlight the need for strategies that address both safety and environmental goals.
| ID | pc605 |
| Presentation | |
| Full Text | |
| Tags |
