The aim of the present research is to investigate driving behavior through the analysis of data collected from connected vehicles. For the purposes of this research, variables related to speed, engine temperature, the anti-lock braking system (ABS), and other driving characteristics were examined. These data were collected over a three-month period to identify different route profiles in terms of driver behavior. The K-Means clustering method was applied to distinguish patterns of driving behavior, and the classification of trips into three clusters was found to provide satisfactory analytical results. Subsequently, the Random Forest algorithm was employed, using the anti-lock braking system as the dependent variable, to assess the importance of the independent variables. Results indicated that the variable with the highest importance value was the engine oil temperature. Finally, a Binary Logistic Regression was developed to examine the extent to which the independent variables affect the probability of ABS activation; revealing that engine activation is the most influential predictor. Overall, the findings highlight the effectiveness of combining connected vehicle data with machine learning techniques to support data-driven road safety analysis and decision-making.
| ID | pc685 |
| Presentation | |
| Full Text | |
| Tags |
