Designing Scalable Interaction Between Autonomous Vehicles and Pedestrians

Abstract

Transportation research increasingly focuses on the interactions between autonomous vehicles (AVs) and pedestrians, with the goal of seamlessly integrating AVs into urban traffic. Engineering research has primarily concentrated on developing methods to understand and predict pedestrian intentions. Meanwhile, interaction design research has examined key elements of AV–pedestrian communication, particularly through external Human–Machine Interfaces (eHMIs).

To date, research in this area has largely provided foundational insights by evaluating eHMIs in scenarios involving a single AV and pedestrian. However, there has been limited exploration of more complex environments featuring multiple AVs and pedestrians, scenarios that are more representative of real-world settings.

Therefore, the aim of this doctoral research is two-fold: First, to explore how eHMIs can be designed to support scalable AV–pedestrian interaction; and second, to generate new insights into the effective use of Virtual Reality (VR) in multi-entity scenarios. To achieve these aims, this research adopts a two-pronged methodology. The first component involves a literature synthesis through scoping reviews, aimed at identifying key challenges in AV–pedestrian interaction and the current state of VR pedestrian simulators. The second component is a research through design approach, characterised by iterative cycles of design, prototyping, and empirical evaluation using VR simulations.

This thesis contributes to the field of Human–Computer Interaction in two key areas: Firstly, the research systematically identifies and addresses the scalability challenges of eHMIs, offering high-level design directions, novel eHMI concepts, and empirically validated design recommendations. Secondly, it introduces methodological advancements by improving the application of VR as a tool for prototyping and evaluation, aimed at supporting the research and development of scalable eHMI solutions.