A Two-Stage Behavioral Model Considering Vehicle Motion Fluctuations for Decision-Making During Lane Changes in Diverging Areas

Optimizing diverging areas is critical for traffic safety and operations management. Inherently, diverging areas require lane changes (i.e., mandatory lane changes), which can cause significant disruptions to local traffic flow. To describe a realistic and detailed diverging area traffic flow, this study developed a two-stage lane-change decision-making behavioral model considering vehicle motion fluctuations. Firstly, according to the geometric characteristics of the diverging area and the need for diverging vehicles to change lanes, a lane-changing rule combining time-headway and distance-headway was established. Meanwhile, a lane-changing probability expression reflecting vehicle movement characteristics and remaining distance was defined. Secondly, the speed fluctuation sensitivity coefficient and distance sensitivity coefficient of the proposed model were calibrated based on real-world mandatory lane-changing observations. Finally, numerical simulations were used to evaluate and verify the proposed model. The results indicate that: 1) macroscopically, the simulated traffic is consistent with the observed traffic in terms of density, flow and speed, and the time–space diagram reflects the similarity of the traffic state changes; 2) microscopically, the section velocity distribution and diversion location distribution indirectly express the driving decision and motion adjustment process of vehicles. This study contributes to in-depth understanding of the traffic flow evolution in diverging areas and facilitates mandatory lane-change modeling.