Accurately modeling microscopic driving behaviors under complex road geometries relies heavily on high-fidelity trajectory datasets. Type-A short weaving segments are typical traffic bottlenecks characterized by intense vehicle conflicts and coexisting discretionary/mandatory lane-changing behaviors, and have therefore long attracted broad. However, no open-source dataset specifically designed for such areas is available, which greatly restricts the validation of microscopic models in these extreme environments. To address this gap, this paper intro-duces U-EASWS, a UAV-based vehicle trajectory dataset for urban expressway Type-A short weaving segments, enriched with comprehensive interaction and risk semantic annotations across seven typical scenarios in Changchun, China. Processed via a standardized automated pipeline, U-EASWS provides three core components: (i) high-fidelity continuous trajectories extracted via a fine-tuned YOLOv8 model and reconstructed using PCHIP interpolation a…
