eHMI for All - Investigating the Effect of External Communication of Automated Vehicles on Pedestrians, Manual Drivers, and Cyclists in Virtual Reality Conference: Proceedings of the 2026 CHI Conference on Human Factors in Computing Systems; April 13–17, 2026; Barcelona, Spain Proceedings of the 202

Mark Colley, Simon Kopp, Debargha Dey, Pascal Jansen, Enrico Rukzio · 2026

Read eHMI for All - Investigating the Effect of External Communication of Automated Vehicles on Pedestrians, Manual Drivers, and Cyclists in Virtual Reality Conference: Proceedings of the 2026 CHI Conference on Human Factors in Computing Systems; April 13–17, 2026; Barcelona, Spain Proceedings of the 202 by Mark Colley, Simon Kopp, Debargha Dey, Pascal Jansen, Enrico Rukzio (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2601-19440.txt.

Introduction

Mark Colley and four co-authors test how self driving cars should signal the people around them. With no human driver to wave or nod, automated vehicles may need external displays to talk to pedestrians, drivers, and cyclists. Past work mostly studied one kind of road user at a time.

The team tests such external interfaces for all three groups at once, inside virtual reality. VR lets them stage risky crossings safely and repeat them exactly. Signals that read clearly to a walker and a cyclist in the same scene are what real streets will demand.

Catalog record

CREATORMark Colley, Simon Kopp, Debargha Dey, Pascal Jansen, Enrico Rukzio
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE68 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2601-19440

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

eHMI for All - Investigating the Effect of External Communication of Automated Vehicles on Pedestrians, Manual Drivers, and Cyclists in Virtual Reality Conference: Proceedings of the 2026 CHI Conference on Human Factors in Computing Systems; April 13–17, 2026; Barcelona, Spain Proceedings of the 2026 CHI Conference on Human Factors in Computing Systems (CHI ’26), April 13–17, 2026, Barcelona, Spain DOI: 10.1145/3772318.3790585 ISBN: 979-8-4007-2278-3/2026/04 CCS: Human-centered computing Empirical studies in HCI Mark Colley, Simon Kopp, Debargha Dey, Pascal Jansen, Enrico Rukzio Abstract . With automated vehicles (AVs), the absence of a human operator could necessitate external Human-Machine Interfaces (eHMIs) to communicate with other road users. Existing research primarily focuses on pedestrian-AV interactions, with limited attention given to other road users, such as cyclists and drivers of manually driven vehicles. So far, no studies have compared the effects of eHMIs across these three road user roles. Therefore, we conducted a within-subjects virtual reality experiment (N=40), evaluating the subjective and objective impact of an eHMI communicating the AV’s intention to pedestrians, cyclists, and drivers under various levels of distraction (no distraction, visual noise, interference). eHMIs positively influenced safety perceptions, trust, perceived usefulness, and mental demand across all roles. While distraction and road user roles showed significant main effects, interaction effects were only observed in perceived usability. Thus, a unified eHMI design is effective, facilitating the standardization and broader adoption of eHMIs in diverse traffic. † † cc-license: by 1. Introduction Automated vehicles (AVs) with an SAE (Society of Automobile Engineers) level ( 84 ) of 4 or 5 will fundamentally transform traffic ( 39 ) and interactions within traffic environments ( 20 ; 29 ; 68 ) . As no human operator needs to be present to communicate with other road users in situations with uncertainty (e.g., regarding right of way), AVs must be equipped with the ability to communicate with other road users in ambiguous situations. This is done through external Human-Machine Interfaces, or short, eHMIs. Previous work proposed numerous eHMIs. For example, displays on windshields ( 31 ) , LED strips ( 41 ; 69 ; 18 ) , anthropomorphic or animal-related communication ( 67 ) , projections ( 2 ; 73 ) , and implicit communication such as movement patterns ( 94 ) . External devices like smartphones ( 50 ) and enhanced infrastructure have also been considered ( 83 ) . Furthermore, studies have already explored different eHMI concepts and designs, like text-based designs ( 9 ) that communicate vehicle status ( 38 ) or intention ( 36 ) . Studies have shown mainly positive effects of eHMIs on trust, clarity, hedonic qualities, and pedestrian crossing behaviors. However, most of them only consider communication between one pedestrian and one AV ( 24 ; 25 ) . Only relatively little work has been conducted with cyclists ( 52 ; 40 ) or drivers of ordinary, manually driven vehicles (MDVs) ( 17 ; 77 ; 78 ) . No work so far has evaluated the effects of eHMIs on all three road users comparatively. Furthermore, it is important to ascertain if a chosen eHMI can function effectively across different kinds of road users to prevent the situation that an AV is equipped with multiple kinds of eHMIs, each catering to a different road user. This can lead to a busy interface, which can, in turn, cause information overload, leading to detrimental effects in the user interface. It is therefore crucial to explore if a holistic approach to eHMI design can work- i.e., whether a unified eHMI is able to perform adequately across different kinds of road users. In this work, we address this gap by employing one unified scene and one eHMI design integrating simulators for MDVs, pedestrians, and cyclists. Additionally, more comprehensive studies that scale up eHMI experiments are required for the results and subsequent insights to be more externally valid. 23 have identified over 38 factors influencing pedestrian crossing decisions, including physical context, dynamic factors, traffic characteristics, social factors, demographics, abilities, and personal characteristics. While some of these factors are dependent on personal, cultural, or national contexts (e.g., demographics, social norms, and law compliance), others relate directly to the external communication of AVs. Therefore, in this work, we also study the influence of distractions (visual noise) or interference (the trajectories of the AV and other traffic intersect, thereby causing an interaction) ( 76 ; 70 ) . We conducted a within-subjects virtual reality (VR) study with N=40 participants to evaluate the effects of eHMIs (no eHMI and an intention-based eHMI) on the participants being in the role of three different road users (pedestrian, driver, cyclist) in scenarios with different levels of distraction (no distraction, visual noise, and interference). In this study with multiple AVs, MDVs, and other pedestrians, eHMIs had positive effects. In these scenarios, the participants generally felt safer, trusted the AVs more, assessed the system as more useful, and felt less mentally demanded. While we also found that distraction and the road user role have significant main effects, we did not find interaction effects except for usability. As a result, we posit that one unified eHMI concept is applicable for diverse road user roles, potentially significantly improving the standardization and adoption of eHMIs. Contribution Statement: This research offers insights into the effects of external communication of AVs for pedestrians, cyclists, and manual drivers. Results from our VR study with N=40 participants suggest that eHMIs are effective across different road user roles. This highlights the potential for a holistic approach to a unified eHMI design that caters to multiple types of road users. 2. Related Work We ground our work in existing research on pedestrian-vehicle interaction, specifically, crossing behaviors in traffic and the space and impact of eHMIs of AVs on traffic interactions. 2.1. External Communication of Automated Vehicles Current traffic interactions often rely on gestures and eye contact to resolve ambiguities, for example, regarding the right of way ( 75 ) . Despite the infrequent necessity for explicit communication ( 64 ) , eHMIs have been proposed to facilitate communication between AVs and vulnerable road users like pedestrians or cyclists ( 49 ) . Prior work has categorized external communication strategies by modality, message type, and location on the vehicle ( 20 ; 21 ; 29 ) . 20 identified eight message types: Instruction, Command, Advisory, Answer, Historical, Predictive, Question, and Affective. Communication locations include various parts of the vehicle, personal devices, or infrastructure, such as sidewalks, with specific focus areas like the windshield or bumper highlighted as key interaction points ( 35 ) . Key considerations for effective eHMI deployment include the communication relationship dynamics (ranging from one-to-one to many-to-many), ambient noise levels, and the specific road user involved (e.g., pedestrian, cyclist) ( 20 ) . Studies have explored eHMI effectiveness across diverse groups, including children ( 28 ; 10 ) , individuals with vision or mobility impairments ( 22 ; 5 ) , and general pedestrians ( 3 ; 32 ; 68 ; 34 ) , as well as cyclists ( 53 ) . Mostly positive outcomes of eHMIs have been reported. For instance, 31 demonstrated that distance-dependent information could significantly improve pedestrians’ understanding of AV intentions and willingness to cross safely. In contrast, 22

Read the complete text