Content uploaded by Daniel Eisele
Author content
All content in this area was uploaded by Daniel Eisele on Apr 02, 2024
Content may be subject to copyright.
Equipping Vehicles with Novel eHMIs Potentially Changes How Pedestrians Interact with Vehicles Without eHMIs
Abstract
Over the last years, there has been a lively discussion whether (automated) vehicles should be equipped with novel external human-machine-interfaces (eHMIs) in order to facilitate communication with nearby vulnerable road users. This exploratory study investigated whether the introduction of eHMI-equipped vehicles to public traffic potentially influences how pedestrians interact with vehicles without eHMIs. To that goal, our participants specified their willingness to cross in front of vehicles that were either equipped with a frontal brake light eHMI or not in a video-based experiment. Between groups, the quota of eHMI-equipped vehicles in simulated traffic was varied. Our findings show that the quota of vehicles with an eHMI did indeed influence street crossing willingness in front of yielding as well as non-yielding vehicles without an eHMI. Notably, the magnitude and direction of the effect was dependent on the distance between vehicle and pedestrian. Future research on eHMIs should take potential unintended side effects of eHMIs into account.
... In one commentary, it has been argued from a gametheoretic perspective that pedestrian behavior will likely change depending on AV market penetration [52]. Notably, results from a first empirical study that simulated different levels of prevalence of vehicles with an eHMI found large effects on pedestrians' willingness to cross in front of vehicles [22]. Interestingly, the authors observed larger effects on interactions with vehicles without any eHMI than vehicles with eHMIs. ...
... In one group, 79 % of these vehicles were AVs, in the other group 21 %. These rates were based on previous research [22]; the exact rates resulted from practical feasibility within our experimental setup. In order to measure an indicator for the way pedestrians interact with a vehicle, the moment they decided to cross the street in front (or behind) a target vehicle was measured [33]. ...
... The results of our video-based laboratory study suggest that the prevalence of AVs can indeed influence pedestrian-vehicle interaction. This is consistent with the results of a previous study on the prevalence of eHMIs, in which a similar methodology was used [22], and reinforces the idea that the interactive behavior of road users can be significantly influenced by the presence of third-party road users who are not directly involved in an interaction. In the present study, we observed that -averaged over all observed crossings -a high prevalence of AVs led to considerably earlier crossing decisions in front of AVs, but no statistically relevant difference in front of CVs. ...
The proportion of highly automated vehicles in traffic (i.e., the prevalence of AVs) is likely to increase over time. The aim of this study was to investigate whether the prevalence of AVs may influence how pedestrians interact with AVs and with conventional, human-driven vehicles (CVs). A video-based laboratory study was conducted using a two-group mixed design. Participants took the perspective of pedestrians about to cross the road in a situation where AVs (with eHMIs) and CVs were approaching their position. The prevalence of AVs was manipulated between groups (low/high). The participants indicated the moment they decided to cross in front of the vehicles. Our results show that AV prevalence did indeed significantly influence when participants decided to cross. Overall, participants decided to cross earlier in front of the more prevalent vehicle type. Therefore, taking into account the given prevalence of AVs could significantly benefit AVs in predicting pedestrian behavior.
... While empirical findings on the effects of eHMIs have been mixed, a consensus has emerged that they can indeed influence pedestrians' perceptions and behaviors, including trust, willingness to cross, and crossing timing [7,[11][12][13][14][15]. Notably, recent studies have shown that eHMIs can influence vehicle-pedestrian interactions not only when activated (e.g., increase willingness to cross by activating) but also when inactivated (e.g., decrease willingness to cross when inactivated) [7,[15][16][17]. In these studies, participants changed their crossing behavior in front of vehicles with inactive eHMIs (and vehicles without eHMIs) if they had previously been exposed to vehicles with active eHMIs. ...
In recent years, there has been a debate on whether automated vehicles (AVs) should be equipped with novel external human–machine interfaces (eHMIs). Many studies have demonstrated how eHMIs influence pedestrians’ attitudes (e.g., trust in AVs) and behavior when they activate (e.g., encourage crossing by lighting up). However, very little attention has been paid to their effects when they do not activate (e.g., discourage crossing by not lighting up). We conducted a video-based laboratory study with a mixed design to explore the potential of two different eHMI messages to facilitate pedestrian-AV interactions by means of activating or not activating. Our participants watched videos of an approaching AV equipped with either a state eHMI (“I am braking”) or intent eHMI (“I intend to yield to you”) from the perspective of a pedestrian about to cross the road. They indicated when they would initiate crossing and repeatedly rated their trust in the AV. Our results show that the activation of both the state and intent eHMI was effective in communicating the AV’s intent to yield and both eHMIs drew attention to a failure to yield when they did not activate. However, the two eHMIs differed in their potential to mislead pedestrians, as decelerations accompanied by the activation of the state eHMI were repeatedly misinterpreted as an intention to yield. Despite this, user experience ratings did not differ between the eHMIs. Following a failure to yield, trust declined sharply. In subsequent trials, crossing behavior recovered quickly, while trust took longer to recover.
... One vital domain of study involves how intelligent machines can alter human behavior, including interactions between humans -a perspective that has received very little attention in AV-ORU so far [18], although we are certainly aware of the complexity and challenges of automation for humans [4]. The more established field of HRI has, for over a decade, produced empirical evidence showing that human-human interactions can be significantly affected by human-robot interactions and that observing robot-robot interactions can also influence human-robot interactions [13,30,33,54,58]. ...
Substantial resources are being invested in integrating social robots and automated vehicles (AVs) into everyday life. I argue that both social robots and AVs with artificial intelligence (AI) can be considered social actors within the broader category of "intelligent machines". If they are indeed integrated into society on a large scale, the fields of Human-Robot Interaction (HRI) and Automated Vehicles-Other Road User Interaction (AV-ORU) face the challenge of studying their behavior and impact within an emerging hybrid human-machine society. However, both disciplines are in their early stages regarding the exploration of interdependencies between their entities of interest and the complex adaptive systems in which they (will) operate. I propose that meta-scientific insights from systems science provide valuable perspectives to guide the formulation of research questions and the design of future studies across both disciplines.
... However, in both studies they note that effects such as distraction through smartphone usage resulting in a safety maneuver or operating at a very low vehicle speed that is not perceived as dangerous might also have an effect on the results. In contrast to that, the studies [9], [13], [14] indicate that pedestrians do react differently when an approaching vehicle is equipped with an eHMI and that eHMIs have a positive impact on the perceived safety and crossing willingness. However, all studies highlight that the effect of eHMIs on the crossing behavior of pedestrians can only be measured when the approaching vehicle is actually yielding. ...
... Mixed traffic could also become more complex due to new forms of communication that AVs may be equipped with. For example, Eisele et al. (2023) demonstrated that equipping vehicles with eHMIs, which indicated the vehicle's braking, pedestrians were more likely to cross in front of the vehicle. At the same time, it also made individuals feel more uncertain when crossing in front of vehicles without eHMIs. ...
Effects of a frontal brake light (FBL, a potential external human–machine interface for automated vehicles) on participants’ self-reported willingness to cross a vehicle’s path were investigated. In a mixed design online study (vehicles in the experimental group were equipped with FBLs, there were no FBLs in the control group), participants observed videos of a vehicle approaching at different speeds from the perspective of a pedestrian standing at the curb. The vehicles exhibited either yielding behavior (braking onset 55 m or 32 m before standstill in front of the pedestrian ’s position) or non-yielding behavior (approach speed was maintained). Participants specified their willingness to cross the vehicle’s path at different distances. When the vehicle yielded (i.e., FBL was activated), willingness to cross was significantly higher in the experimental group than the control group. Notably, we further observed a significantly lower willingness to cross in the experimental group than the control group when the vehicle did not yield (i.e., FBL was deactivated). Novel external human–machine interfaces might therefore influence the interaction with vehicles not only when they are activated but also when they are deactivated.
Given the rise of automated vehicles from an engineering and technical perspective, there has been increased research interest concerning the Human and Computer Interactions (HCI) between vulnerable road users (VRUs, such as cyclists and pedestrians) and automated vehicles. As with all HCI challenges, clear communication and a common understanding—in this application of shared road usage—is critical in order to reduce conflicts and crashes between the VRUs and automated vehicles. In an effort to solve this communication challenge, various external human–machine interface (eHMI) solutions have been developed and tested across the world. This paper presents a timely critical review of the literature on the communication between automated vehicles and VRUs in shared spaces. Recent developments will be explored and studies analyzing their effectiveness will be presented, including the innovative use of Virtual Reality (VR) for user assessments. This paper provides insight into several gaps in the eHMI literature and directions for future research, including the need to further research eHMI effects on cyclists, investigate the negative effects of eHMIs, and address the technical challenges of eHMI implementation. Furthermore, it has been underlined that there is a lack of research into the use of eHMIs in shared spaces, where the communication and interaction needs differ from conventional roads.
Pedestrian route choice, the process by which individuals decide on their walking path between two locations, is a fundamental problem across disciplines. Because this behaviour is investigated from different conceptual and methodological angles, and because it strongly depends on the environmental context, it is challenging to establish a systematic framework for research. Here, by reviewing previous work, we identify four principles for pedestrian route choice that are relevant across disciplines. First, ‘information perception’ deals with how pedestrians can perceive information selectively and purposely, given the limited available information. Second, ‘information integration’ considers how pedestrians subjectively integrate environmental spatial information into mental representations. Third, ‘responding to information’ is concerned with how pedestrians tend to be attracted and repelled by specific attributes individually and how this can lead to positive or negative feedback loops across many individuals. Fourth ‘decision-making mechanisms' describe how pedestrians trade off the evidence provided by different attributes. How pedestrians perceive, integrate, respond to, and act upon information is not fixed but varies with the context. We give examples for each principle and explain how these principles shape pedestrian choice behaviours. We hope this contribution provides a systematic overview of the field and helps to spark inspiration among specialists.
Introduction:
In current urban traffic, pedestrians attempting to cross the road at un-signalized locations are thought to mostly use implicit communication, such as deceleration cues, to interpret a vehicle's intention to yield. There is less reliance on explicit driver- or vehicle-based messages, such as hand/head movements, or flashing lights/beeping horns. With the impending deployment of Automated Vehicles (AV), especially those at SAE Level 4 and 5, where the driver is no longer in control of the vehicle, there has been a surge in interest in the value of new forms of communication for AVs, for example, via different types of external Human Machine Interfaces (eHMIs). However, there is still much to be understood about how quickly a novel eHMI affects pedestrian crossing decisions, and whether it provides any additional aid, above and beyond implicit/kinematic information from the vehicle. The aim of this between-participant study, funded by the H2020 interACT project, was to investigate how the combination of kinematic information from a vehicle (e.g., Speed and Deceleration), and eHMI designs, play a role in assisting the crossing decision of pedestrians in a cave-based pedestrian simulator.
Method:
Using an existing, well-recognized, message for yielding (Flashing Headlights - FH) as a benchmark, this study also investigated how quickly a novel eHMI (Slow Pulsing Light Band - SPLB) was learned. To investigate the effect of eHMI visibility on crossing decisions, the distance at which each eHMI was perceivable was also measured.
Results:
Results showed that, compared to SPLB, the FH led to earlier crossings during vehicle deceleration, especially at lower approaching speeds, and smaller time gaps. However, although FH was visible earlier than SPLB, this visibility does not appear to be the only reason for earlier crossings, with message familiarity thought to play a role. Participants were found to learn the meaning conveyed by FH relatively quickly, crossing around 1 second earlier in its presence (compared to the no eHMI condition), across the three blocks of trials. On the other hand, it took participants at least one block of 12 trials for the new SPLB signal to affect crossing, which only accelerated crossing initiations by around 200 ms, compared to the no eHMI condition. The role of comprehension, long-term exposure, and familiarity of novel messages in this context is therefore important, if AVs are to provide safe, trustworthy communication messages, which will enhance traffic flow and efficiency.
There is a growing body of research in the field of interaction between automated vehicles and other road users in their vicinity. To facilitate such interactions, researchers and designers have explored designs, and this line of work has yielded several concepts of external Human-Machine Interfaces (eHMI) for vehicles. Literature and media review reveals that the description of interfaces is often lacking in fidelity or details of their functionalities in specific situations, which makes it challenging to understand the originating concepts. There is also a lack of a universal understanding of the various dimensions of a communication interface, which has impeded a consistent and coherent addressal of the different aspects of the functionalities of such interface concepts. In this paper, we present a unified taxonomy that allows a systematic comparison of the eHMI across 18 dimensions, covering their physical characteristics and communication aspects from the perspective of human factors and human-machine interaction. We analyzed and coded 70 eHMI concepts according to this taxonomy to portray the state of the art and highlight the relative maturity of different contributions. The results point to a number of unexplored research areas that could inspire future work. Additionally, we believe that our proposed taxonomy can serve as a checklist for user interface designers and researchers when developing their interfaces.
Objective:
In this article, we investigated the effects of external human-machine interfaces (eHMIs) on pedestrians' crossing intentions.
Background:
Literature suggests that the safety (i.e., not crossing when unsafe) and efficiency (i.e., crossing when safe) of pedestrians' interactions with automated vehicles could increase if automated vehicles display their intention via an eHMI.
Methods:
Twenty-eight participants experienced an urban road environment from a pedestrian's perspective using a head-mounted display. The behavior of approaching vehicles (yielding, nonyielding), vehicle size (small, medium, large), eHMI type (1. baseline without eHMI, 2. front brake lights, 3. Knightrider animation, 4. smiley, 5. text [WALK]), and eHMI timing (early, intermediate, late) were varied. For yielding vehicles, the eHMI changed from a nonyielding to a yielding state, and for nonyielding vehicles, the eHMI remained in its nonyielding state. Participants continuously indicated whether they felt safe to cross using a handheld button, and "feel-safe" percentages were calculated.
Results:
For yielding vehicles, the feel-safe percentages were higher for the front brake lights, Knightrider, smiley, and text, as compared with baseline. For nonyielding vehicles, the feel-safe percentages were equivalent regardless of the presence or type of eHMI, but larger vehicles yielded lower feel-safe percentages. The Text eHMI appeared to require no learning, contrary to the three other eHMIs.
Conclusion:
An eHMI increases the efficiency of pedestrian-AV interactions, and a textual display is regarded as the least ambiguous.
Application:
This research supports the development of automated vehicles that communicate with other road users.
The number of pedestrian casualties in crashes with motorised vehicles is still alarming. Misunderstandings about the other road users’ intentions are certainly one contributory factor. Especially given recent developments in vehicle automation, informing about “vehicle behaviour” and “vehicle intentions” in the absence of any direct interaction between the driver and the outside world is becoming increasingly relevant. A frontal brake light which communicates
that a vehicle is decelerating could be a simple approach to support pedestrians and other road users in the interaction with (potentially automated) motorised vehicles. To assess the effect of a frontal brake light on the identification of vehicle deceleration, we conducted a video based lab experiment. The brake light facilitated the identification of
decelerations considerably. At the same time, the fact that only half of the decelerations were accompanied by the brake light resulted in increased identification times for decelerations in which the frontal brake light was absent compared to a control condition in which none of the decelerations was indicated by such a light. This finding points towards an increasingly conservative approach in the participants’ assessment of deceleration, which could be interpreted as an
indicator for a potential safety effect of the frontal brake light.
In this study, the effect of a frontal brake light (FBL) on children’s willingness to cross the road was investigated. While recent studies have investigated effects of an FBL their samples consisted exclusively of adults. The results and conclusions of these studies may not be applicable to children due to their partially less developed cognitive capabilities. Accordingly, the aim of this study was to investigate the effects of an FBL on children’s willingness to cross a vehicle’s path. In a mixed design simulation study, participants were assigned to an experimental condition (EC), in which vehicles were equipped with an FBL, or to a control condition (CC), in which vehicles were not equipped with an FBL. Children aged six to twelve watched videos from the perspective of a pedestrian standing at the curb. A vehicle with an initial speed of 30 km/h approached either by maintaining the speed or by decelerating. In the latter case, the braking onset was varied (55 m or 32 m from the pedestrian’s position). Participants’ task was to indicate their willingness to cross
the road in front of the vehicle at five different distances. Results show that FBL affected the willingness to cross the road, especially at an early braking onset (i.e., 55 m). If the vehicle decelerated (i.e., the FBL was activated in the EC), the willingness was significantly higher in the EC than in the CC. In case the vehicle maintained its speed (i.e., the FBL was not activated in the EC), the opposite effect appeared. However, children did not exclusively rely on information gained by the FBL, but still paid attention to distance to the vehicle and braking onset. In conclusion, the effects of an FBL on children are similar to the effects on adults identified in prior studies. Nevertheless, more research in complex scenarios is needed to draw general conclusions about the impact of FBL on road safety in differently aged children and adults.
Policymakers recommend that automated vehicles (AVs) display their automated driving status using an external human-machine interface (eHMI). However, previous studies suggest that a status eHMI is associated with overtrust, which might be overcome by an additional yielding intent message. We conducted a video-based laboratory study (N = 67) to investigate pedestrians’ trust and crossing behavior in repeated encounters with AVs. In a 2x2 between-subjects design, we investigated (1) the occurrence of a malfunction (AV failing to yield) and (2) system transparency (status eHMI vs. status+intent eHMI). Results show that during initial encounters, trust gradually increases and crossing onset time decreases. After a malfunction, trust declines but recovers quickly. In the status eHMI group, trust was reduced more, and participants showed 7.3 times higher odds of colliding with the AV as compared to the status+intent group. We conclude that a status eHMI can cause pedestrians to overtrust AVs and advocate additional intent messages.
This paper discusses whether the knowledge of the driving mode of an approaching vehicle (manual vs. automated) influences pedestrians’ decisions while crossing a street. Additionally, the paper explores how different appearances and driving behaviours of vehicles interact with driving mode in affecting pedestrians’ road-crossing behaviours. In a video-based experiment with sixty participants, two vehicles with different appearances (a BMW 3 and a Renault Twizy) were presented as either manually-driven or automated vehicles. Both vehicles displayed either yielding or non-yielding behaviour on a straight road devoid of other traffic. Participants were asked to indicate whether they would cross the street in front of the approaching vehicle, at different distances ranging from 45 m to 1.5 m. The results showed that there was no significant influence of the knowledge of the driving mode (manually-driven vs automated) on pedestrians’ willingness to cross the street at any distance. The vehicle’s behaviour (whether it is maintaining speed or yielding) played a dominant role in pedestrians’ decision to cross a road, and this was similar for both modes and both vehicles, at all distances. However, results suggested that in situations and at distances when the intent of the vehicle was not fully clear by the behaviour of the car alone, there were differences between the two vehicles at certain distances, which could be attributed to the differences in their appearance such as size, aggressiveness and novelty. A futuristic-looking vehicle inspired less confidence in road-crossing situations compared to an ordinary-looking vehicle. Additionally, a novel and futuristic-looking vehicle appeared to make it easier for people to believe that it is an automated vehicle. We conclude by discussing design implications for the development of external HMIs automated vehicles.