Driving style characteristics based lane-changing intention recognition research for truck drivers near highway ramps.

Objective: The research aims to analyze the driving styles and lane-changing intentions of truck drivers near the highway ramps.

Methods: Using principal component analysis (PCA), three principal components were selected for cluster analysis, examining driving styles from the perspectives of risk tolerance, longitudinal, and lateral driving characteristics. An intention recognition model for lane-changing was developed and trained, and its validity was verified with High-D dataset.

Results: The proposed model in this study achieves an accuracy of 93.7% and an F1 score of 0.891, demonstrating its excellent performance in precision-related metrics. Moreover, the study compares the differences in driving styles and lane-changing intentions between truck drivers and sedan drivers.

Conclusions: The main conclusions are as follows: the lane-changing process consists of two stages: intention and execution. Driving style is a critical factor in the establishment of lane-changing intention models. Four seconds is a proper time window for lane-changing intention prediction. The lane-changing behavior characteristics of truck drivers differ significantly from those of sedan drivers. The study results improve the understanding of truck lane-changing behavior near highway ramps, and they also help to figure out the safety mechanisms in the future human-vehicle cooperative traffic scenarios.