Correlation Analysis of Drivers’ Natural Driving Behavior Based on Kernel Density Estimation

Features
Authors Abstract
Content
To investigate the interplay between driver handling behaviors, this article collects data on vehicle kinematic parameters characterizing driver handling characteristics under natural driving, estimates the probability density curves of the parameters using the kernel density method, and fits the curve equations. On this basis, a percentile correlation analysis was performed between the parameters to obtain the influence relationship between the handling behaviors. The results show that longitudinal maneuvers are frequent and intense in the 0–10 km/h speed range, lateral maneuvers are more intense in the 10–30 km/h speed range, and the interaction between longitudinal and lateral maneuvers is more intense in the acceleration phase. This study enriches the natural driving dataset and illustrates the correlation of driving behavior under natural driving, providing a theoretical and data basis for the development of driver-oriented intelligent driving technologies.
Meta TagsDetails
DOI
https://doi.org/10.4271/13-04-02-0012
Pages
24
Citation
Sun, T., Hu, H., Cai, R., Yu, T. et al., "Correlation Analysis of Drivers’ Natural Driving Behavior Based on Kernel Density Estimation," SAE Int. J. Sust. Trans., Energy, Env., & Policy 4(2):189-212, 2023, https://doi.org/10.4271/13-04-02-0012.
Additional Details
Publisher
Published
Jun 9, 2023
Product Code
13-04-02-0012
Content Type
Journal Article
Language
English