Two-dimensional dynamic analysis of vehicle sideway lane changing based on size-dependent driving risk potential field.

Abstract

This paper investigates sideway lane changing in complex traffic scenarios around large-sized vehicles from a microscopic perspective by incorporating vehicle dimensions into the driving risk potential field (DRPF) model. Large-sized vehicles generate a stronger DRPF, which extends into adjacent lanes, enabling effective two-dimensional analysis of traffic behaviors. Through the analysis of the DRPF and the associated driving risk force field (DRFF), sideway lane changing are categorized into three categories: rear-side lane changing, lateral-side lane changing, and front-side lane changing. Statistical results reveal that lateral-side and rear-side lane-changing events account for the majority of sideway lane-changing occurrences, with spatial distribution density indicating that the rear half of the vehicle body is a high-frequency zone for such maneuvers. Furthermore, event-level parameter calibration and trajectory fitting were conducted for three representative sideway lane-changing events, representing rear-side, lateral-side, and front-side interaction scenarios, respectively. For these three cases, the simulated trajectories captured the main features of the observed two-dimensional vehicle motions, with relatively small positional errors. The DRPF model extends traditional one-dimensional traffic behavior analysis to a more realistic two-dimensional framework, validating its its capability and goodness of fit in studying complex traffic behaviors such as sideway lane changing. This advancement highlights its potential for broader applications in cutting-edge traffic research, providing a basis for future investigations of trajectory prediction and autonomous driving applications.