Optimizing Lane Transition Zone Length and Minimum Entry Clearance Distance in Merging Areas With Longitudinal Traffic Separator for Expressway Reconstruction and Expansion

Merging‐zone safety remains a critical challenge in transportation engineering due to frequent severe accidents and insufficient research on longitudinal traffic separator (LTS) configurations and the limited availability of evidence‐based guidance for LTS openings in interchange merging areas. This study addresses this gap by developing a physics‐informed computational framework to optimize lane transition zone (LTZ) length and minimum entry clearance distance (MECD). Specifically, (i) UAV videos collected at three Xi’an Ring Expressway interchanges (438 vehicles) are processed using YOLOv5‐based detection to extract vehicle trajectories; (ii) lane‐changing kinematics are characterized using a modified hyperbolic‐tangent trajectory model, and a gap‐acceptance‐based waiting component is incorporated to compute the required lane‐changing and waiting distances for LTZ/MECD design; and (iii) the recommended values are evaluated via controlled simulation experiments across multiple design speeds and lane configurations using delay and conflict‐rate metrics. Through the empirical analysis of 438 vehicles at three Xi’an Ring Expressway interchanges using a UAV and YOLOv5 to extract trajectories, the framework achieves dual safety and efficiency outcomes. It maintains conflict rates below the critical threshold of 0.76 conflicts/veh·km while increasing throughput by 50%–78% in simulated scenarios. The derived design relationships—LTZ ∝ V 2 /C (where V is the design speed and C is the capacity) and MECD = f (V, C, n) (where n is the number of adjacent lanes)—establish quantitative geometric optimization strategies for improving LTS configurations and provide implementable geometric guidance for expressway reconstruction and expansion applications.