A Video-Based, Eye-Tracking Study to Investigate the Effect of eHMI Modalities and Locations on Pedestrian–Automated Vehicle Interaction

Abstract

Numerous studies have emerged on the external human–machine interface (eHMI) to facilitate the communication between automated vehicles (AVs) and other road users. However, it remains to be determined which eHMI modality and location are proper for the pedestrian–AV interaction. Therefore, a video-based, eye-tracking study was performed to investigate how pedestrians responded to AVs with eHMIs in different modalities (flashing text, smiley, light band, sweeping pedestrian icon, arrow, and light bar) and locations (grill, windshield, and roof). Moreover, the effects of pedestrian-related factors (e.g., gender, sensation-seeking level, and traffic accident involvement) were also included and evaluated. The dependent variables included pedestrians’ clarity-rating scores towards these eHMI concepts, road-crossing decision time, and gaze-based metrics (e.g., fixation counts, dwell time, and first fixation duration). The results showed that the text, icon, and arrow-based eHMIs resulted in the shortest decision time, highest clarity scores, and centralized visual attention. The light strip-based eHMIs yielded no significant decrease in decision time yet longer fixation time, indicating difficulties in comprehension of their meaning without learning. The eHMI location had no effect on pedestrians’ decision time but a substantial influence on their visual searching strategy, with a roof eHMI contradicting pedestrians’ inherent scanning pattern. These findings provide implications for the standardized design of future eHMIs.