| Summary: | This dissertation explores the distinctions in different vehicular behavior during lane-changing scenarios by analysing the public transportation datasets highD. Through detailed data analysis, we describe the differences in dynamics across these scenarios. Building on this foundation, we have developed a dectction model that predicts and categorizes cut-in maneuvers with precision. Despite extensive studies on general lane-changing, the distinct nature and implications of cut-in maneuvers have been insufficiently explored. This study utilizes the comprehensive highD dataset to conduct a detailed comparative analysis of cut-ins versus other lane-changing actions. We extract and categorize lane-changing events, applying gap-based rules to distinctly identify cut-in maneuvers. The research develops and employs a set of performance metrics to assess and compare the driving characteristics inherent in these two categories of lane changes. Employing the Wilcoxon rank-sum test, our analysis reveals significant behavioral differences in these two classification. Furthermore, our study ventures into predictive modeling of cut-in behavior, aiming to enhance the understanding and predictability of its maneuvers. Our findings highlight the critical need for specialized focus on cut-in maneuvers, offering valuable insights for future research in vehicular dynamics and the advancement of autonomous driving systems.
|
|---|