Research on lateral driving style classification based on the NGSIM database

In the context of intelligent and connected transportation systems, intelligent driving systems can effectively reduce driver workload, enhance passenger comfort, ensure driving safety, and improve road traffic efficiency. This study addresses the high complexity of the lane-change process in driver-operated vehicles by analyzing the trajectory data of the NGSIM (Next Generation Simulation) I-80 highway segment from 04:00 to 04:15. We propose an adaptive sine cosine–Levy flight– quantum particle swarm optimization algorithm fused with Support Vector Machine (ASL-QPSO-SVM) to reclassify the driving characteristics initially obtained via K-means clustering. A driving behavior classification model is then established. When a driver intends to change lanes, the intelligent driver assistance system of the human-machine collaborative vehicle can simulate brain-like behavior to accurately identify the driving style of both the vehicle and surrounding vehicles. This allows for precise judgment of the driver’s tendencies, ensuring that the lane-change strategy is both more efficient and safer. This approach holds significant importance for improving the overall safety and efficiency of autonomous driving systems in mixed traffic environments.