Platoon control and external human–machine interfaces: innovations in pedestrian–autonomous vehicle interactions

Abstract

At uncontrolled mid-block crosswalks, autonomous vehicles (AVs) are legally and safely obligated to yield to pedestrians, potentially exacerbating traffic congestion. This study explores integrating AV platoon control with eHMIs to examine its long-term potential for enhancing traffic flow safety and efficiency. A novel agent-based pedestrian-vehicle interaction framework is developed to evaluate the proposed strategy, accurately reproducing the heterogeneous behaviors of pedestrians and vehicles, such as pedestrians' risk tolerance, their responses to eHMIs, and vehicles' yielding mechanisms. Additionally, a platoon formation module is proposed to achieve the desired speed and inter-vehicular spacings from arbitrary initial condition. Simulation results indicate that implementing the proposed strategy reduced the conflict rate by 10%, increased the yielding rate by 11%, and decreased aggregate delay by 34.6% in typical flow conditions. The sensitivity analysis reveals that the effectiveness of the proposed strategy in improving traffic safety and efficiency increases with higher pedestrian and vehicle flow rates.