A novel merging strategy model considering the remaining distance in the acceleration lane

Abstract

The driver of a merging vehicle must account for both the lane‐changing risk and the remaining distance in the acceleration lane () during a merging decision‐making process. To investigate the impact of on merge decisions, a typical freeway merging section was selected and monitored with a millimetre radar and a high‐resolution digital camera, which were mounted on the guardrail in the gore area. More than 2000 merging vehicles were captured during the data collection process. The effects of the surrounding vehicles on the merging behaviour were analyzed, and a merging strategy model considering that was based on the random forest algorithm was constructed. The results show that the following vehicle in the target lane is the main factor that affects the merging behaviour of the merging vehicle. When the decision time window was set to 0.6 s, the proposed merging decision model could distinguish ‘Merge’ events and ‘Wait’ events with accuracies of 97.2% and 89.4%, respectively. The overall accuracy of the model was 94.9%, which was 3.9% higher than for a corresponding merging decision model that excluded influence. The proposed merging decision model can aid merging processes and give cues for human‐like merge decisions of automated vehicles.