Vehicles equipped with an Automated Driving System (ADS) have the potential to significantly reduce road collisions. To enable widespread adoption of ADSs, rigorous safety assessment is essential. Valuable insights for ADS safety validation can be gained by simulating scenarios across a broad range of feature variations. A common challenge in simulating these scenarios is known as the curse of dimensionality, where increasing the number of scenario features requires a near-infinite number of simulations to cover all variations. This issue of complexity presents a need for reducing scenario features. Most related work focuses on identifying important scenario features, while few evaluate how reducing these features impacts ADS failure estimation. The present study aims to address this gap by employing a wide range of feature reduction methods and assessing their effect on ADS failure estimation. Previous research generated datasets for three distinct scenario categories by performing virtual simulations using driver reference models on real-world data. In the present work, the machine learning classifiers such as extreme gradient boosting and random forest are applied to this data for predicting ADS failures. Ten dimensionality reduction techniques, including both feature selection and transformation approaches, are employed to reduce the scenario feature set. The optimal reduced feature set is selected based on classification performance measured by the area under the precision-recall curve. To assess the impact on ADS failure estimation, results are compared against those obtained with the full set of features. The findings indicate that reliable ADS failure estimates can be maintained, and even significantly improved, after substantially reducing the number of scenario features. By reducing scenario features, fewer virtual simulations may be required to reliably estimate ADS failures, which may enable more efficient scenario-based ADS safety assessment. Additionally, this study may offer guidance on selecting suitable dimensionality reduction techniques for scenario-based ADS safety assessment.

This study examines how cyclists’ anticipatory yielding behavior is influenced by spatial and social contexts. While previous research has shown that cyclists’ yielding behavior is shaped by environmental context, a controlled study isolating the effects of specific spatial and social factors would be a valuable addition. Two controlled video simulation experiments were conducted (total N = 67), in which participants viewed traffic scenarios from a cyclist’s perspective. In each video, a pedestrian crossed, and participants indicated their intended anticipatory behavior in the presented scenario by pressing Continue or Brake. The experiments manipulated kinematics (time-to-arrival and distance gap), pedestrian density, and presence or absence of a dedicated cycling lane. Participants selected braking less often and responded more slowly in low pedestrian-density environments and when a dedicated cycling lane was present. Greater time-to-arrival and distance gaps to the pedestrian (i.e., kinematics) were also associated with fewer selected braking responses and increased response times. Pedestrian density had a stronger influence in shared spaces, while time gap effects weakened under low-density conditions. Together, these findings suggest that participants’ reported yielding responses are influenced by spatial (dedicated cycling lane, kinematics) and social factors (pedestrian density). Implications for infrastructure design are discussed.

Efficient attention deployment in visual search is limited by human visual memory, yet this limitation can be offset by exploiting the environment's structure. This paper introduces a computational cognitive model that simulates how the human visual system uses visual hierarchies to prevent refixations in sequential attention deployment. The model adopts computational rationality, positing behaviors as adaptations to cognitive constraints and environmental structures. In contrast to earlier models that predict search performance for hierarchical information, our model does not include predefined assumptions about particular search strategies. Instead, our model's search strategy emerges as a result of adapting to the environment through reinforcement learning algorithms. In an experiment with human participants we test the model's prediction that structured environments reduce visual search times compared to random tasks. Our model's predictions correspond well with human search performance across various set sizes for both structured and unstructured visual layouts. Our work improves understanding of the adaptive nature of visual search in hierarchically structured environments and informs the design of optimized search spaces.

This workshop aims to bring together researchers and practitioners interested in the future of work and the domains of mobility and transportation. Transportation used to be a means to get to and from work. However, three trends give grounds to reconsider the intersection of work with mobility and transportation: (1) People are changing where theywork, including when on the move (e.g., on the train, plane, or even in their car and while walking), (2) Those for whom transportation was part of their job are experiencing changes due to the availability of more technology (e.g., cab/truck/train drivers, first responders, delivery couriers), and (3) People who work to ensure that others keep moving (e.g., train planners, urban planners) face more diverse types of transportation and technical support that changes theirwork. In thisworkshop, we will exchange ideas with participants interested in the intersection of (the future of) work and mobility and transportation. The goal is to stimulate community-driven bottom-up initiatives about how (the future of) work reshapes the mobility and transportation domain.

This paper investigates whether empirical findings on how humans evaluate arguments in reinstatement cases support the ‘fewer attackers is better’ principle, incorporated in many current gradual notions of argument acceptability. Through three variations of an experiment, we find that (1) earlier findings that reinstated arguments are rated lower than when presented alone are replicated, (2) ratings at the reinstated stage are similar if all arguments are presented at once, compared to sequentially, and (3) ratings are overall higher if participants are provided with the relevant theory, while still instantiating imperfect reinstatement. We conclude that these findings could at best support a more specific principle ‘being unattacked is better than attacked’, but alternative explanations cannot yet be ruled out. More generally, we highlight the danger that experimenters in reasoning experiments interpret examples differently from humans. Finally, we argue that more justification is needed on why, and how, empirical findings on how humans argue can be relevant for normative models of argumentation.

This study investigates the effect of using a systematic viewing protocol for scanning closed-circuit television (CCTV) during a simulation of remote nautical object (bridge/lock) control. For nautical object control, systematic viewing is assumed to mitigate the risk of observer errors, such as missing a road user in CCTV streams. However, previous research has reported mixed results on the benefits of systematic viewing for performance. A total of 42 professional operators were asked to operate a bridge control simulator where, unknown to the operators, critical events had to be detected. Half of the group received protocol instructions, the other half did not. The protocol group showed significantly longer dwell times and higher coverage of protocol-related CCTV areas than the no-protocol group. Critical event detection rates were identical for the two groups. While timing of the first fixation on the critical events did not significantly differ between groups, the no-protocol group responded significantly faster to the events. Although protocol application did not improve or impair detection performance, the combined results suggest that the protocol group took more time to scan the scene before acting. As nautical object control prioritizes safety over speed, increased dwell time and coverage are beneficial.

The urban digital twin (UDT) is derived from the original digital-twin concept of a representation of physical assets. This has left the social component of the city underrepresented in UDTs. Here, we discuss what this means for the current maturity stage of UDTs and why better representing human behaviour in UDTs may diversify possibilities to support different types of planning. We contemplate operationalizing the representation of human behaviour by means of agent-based models (ABMs) integrated with UDTs and illustrate this with two concrete examples of simulating stress and safety perception in public spaces. One example shows the idea of the UDT as a live data repository for ABMs, with the ABM adding dynamism, and the other of live feedback between the city, the ABM and UDT. We discuss several epistemological, conceptual, technical, and ethical challenges that may be involved in this integration. We conclude with a future agenda to promote (1) the abandonment of the vision of a UDT as the highly detailed mirror of the city, (2) UDTs fit for sectoral (strategic) in addition to operational planning, (3) the inclusion of behavioural and social processes in UDTs by incorporating ABMs, (4) a culture of cumulative research using structured guided frameworks and reusable building blocks, (5) ABMs with explicit purposes to allow fit-for-purpose selection in UDTs, and (6) explicitly addressing epistemic, normative, and moral responsibilities. Thus, though including agents may at some point be a solution for the (currently lacking) perspective on the role of humans in shaping and being shaped by the city, several reconsiderations in the UDT and ABM communities need to take place first.

We compared how different levels of cognitive load affect frontal P3 (fP3) Event-Related Potential (ERP) to novel sounds. Previous studies demonstrated the predictive value of the probe-elicited frontal P3 (fP3) ERP for subsequent detection failures. They also demonstrated how fP3 is reduced when performing visual and/or manual and/or cognitively demanding tasks. These results are consistent with fP3 indexing orienting to novels or, more neutrally: susceptibility. Here, we tested how fP3 is affected by a threefold variation of cognitive load induced by the verb (generation) task. Participants heard a noun and either listened to it, repeated it, or generated a semantically related verb. These conditions were manipulated between groups. One group (N = 16) experienced the listen and repeat condition; the other group (N = 16) experienced the listen and generate condition. When fP3 was probed 0 or 200 ms after noun offset, it was reduced (relative to no noun) only while repeating or generating, not while listening. An additional probe-elicited ERP was identified as novelty-related negativity, and its contaminating influence on fP3 estimation accounted for by a novel vector-filter procedure. We conclude that cognitive load does not affect fP3-indexed susceptibility. Instead, fP3-indexed susceptibility is affected by presentation of the stimulus, with the most pronounced effect in conditions where a vocal response is needed (i.e., repeat or generate, but not listen), independent of the complexity of the response.

This paper reports a qualitative exploration of gaze strategies during closed-circuit television (CCTV) tasks in remote nautical object control of a lock. Previous research has not examined gaze strategies in scenarios where systems, such as nautical objects, are operated remotely using CCTV. As contextual factors matter in nautical object control, a qualitative approach was necessary to uncover domain-specific terminology and insights into gaze strategies. We recorded eye gaze from professional lock operators and then conducted semi-structured interviews with domain experts to assess these recordings. Thematic analysis revealed that experts were able to identify (features of) gaze strategies but did not share the same terminology. Based on this analysis we defined four strategies: anticipating, verifying, overview, and movement-directed gazes (RQ1). All strategies, except movement-directed gaze, were observed consistently across operators (RQ2), with verifying gaze aligning with task steps in a predefined protocol (RQ3). More generally, our classification framework from thematic analysis could help to systematically define and verify gaze strategies based on domain and task features across various CCTV working contexts. For nautical object control, the framework can be instrumental in interpreting and verifying future (quantitative) eye tracking results and informing instructional procedures.

We discuss the state-of-the-art and future directions of the development, evaluation, and application of computational cognitive models for human-automated vehicle interaction. The capabilities of automated vehicles are rapidly increasing and changing human interaction with and around the vehicle. Yet, at the same time, fully automated vehicles that do not require human interaction are not available. Therefore, systems are needed in which the human and the vehicle interact together. We discuss how computational cognitive models that can describe, predict, and/or anticipate human behavior and thought can play a crucial role in this regard. Such research comes from many different disciplines including cognitive science, human-computer interaction, human factors, transportation research, and artificial intelligence. This special issue brings together state-of-the-art research from these fields. We identify four broader directions for future research: (1) to continue Allen Newell's research agenda for cognitive modeling, but now apply it to the field of human-automated vehicle interaction; (2) to move from isolated theory-slicing to integrated theories, (3) to consider cognitive models both for analysis of interaction and for use in embedded systems; (4) to move from models that mostly describe to models that can predict.

This paper reports results from a high-fidelity driving simulator study (N=215) about a head-up display (HUD) that conveys a conditional automated vehicle's dynamic "uncertainty" about the current situation while fallback drivers watch entertaining videos. We compared (between-group) three design interventions: display (a bar visualisation of uncertainty close to the video), interruption (interrupting the video during uncertain situations), and combination (a combination of both), against a baseline (video-only). We visualised eye-tracking data to conduct a heatmap analysis of the four groups' gaze behaviour over time. We found interruptions initiated a phase during which participants interleaved their attention between monitoring and entertainment. This improved monitoring behaviour was more pronounced in combination compared to interruption, suggesting pre-warning interruptions have positive effects. The same addition had negative effects without interruptions (comparing baseline display). Intermittent interruptions may have safety benefits over placing additional peripheral displays without compromising usability.

We review the state of open science and the perspectives on open data sharing within the automotive user research community. Openness and transparency are critical not only for judging the quality of empirical research, but also for accelerating scientific progress and promoting an inclusive scientific community. However, there is little documentation of these aspects within the automotive user research community. To address this, we report two studies that identify (1) community perspectives on motivators and barriers to data sharing, and (2) how openness and transparency have changed in papers published at AutomotiveUI over the past 5 years. We show that while open science is valued by the community and openness and transparency have improved, overall compliance is low. The most common barriers are legal constraints and confidentiality concerns. Although research published at AutomotiveUI relies more on quantitative methods than research published at CHI, openness and transparency are not as well established. Based on our findings, we provide suggestions for improving openness and transparency, arguing that the motivators for open science must outweigh the barriers. All supporting materials are freely available at: https://osf.io/zdpek/

In this study, we focus on different strategies drivers use in terms of interleaving between driving and non-driving related tasks (NDRT) while taking back control from automated driving. We conducted two driving simulator experiments to examine how different cognitive demands of texting, priorities, and takeover time budgets affect drivers’ takeover strategies. We also evaluated how different takeover strategies affect takeover performance. We found that the choice of takeover strategy was influenced by the priority and takeover time budget but not by the cognitive demand of the NDRT. The takeover strategy did not have any effect on takeover quality or NDRT engagement but influenced takeover timing.

This paper presents PREDICTOR (PREDICting Take-Over Response time): an interactive open-source research software tool to predict the timing of various stages of a transition of control, or take-over, in semi-automated driving. Although previous work has investigated extensively what factors affect the minimum time needed for a successful take-over by the driver, less is known about how specific stages within the take-over process are affected by those factors. PREDICTOR applies a theoretical framework that describes the take-over process as interruption handling through a series of stages. It then ties this theory to a database that summarizes results from previous take-over studies. PREDICTOR can be used to interactively predict through simulation how specific human factors (e.g., alert modality, alert onset time) impact four distinct stages of the take-over response process. The tool simulates and visualizes expected reaction time distributions for each stage of the take-over process. The use of distributions also highlights the likelihood of an accident – as long responses (“outliers”) are quantifiable. Moreover, it can help understand at which stage drivers might take relatively longer or shorter, and which stages are most impacted by a specific factor (e.g., alert modality). PREDICTOR also allows users to add their own data, and to define their own dependent variables for analysis. As a tool that allows exploration of various scenarios, PREDICTOR can aid in the prediction and analysis of potential future accidents.

Objective We experimentally test the effect of cognitive load on auditory susceptibility during automated driving. Background In automated vehicles, auditory alerts are frequently used to request human intervention. To ensure safe operation, human drivers need to be susceptible to auditory information. Previous work found reduced susceptibility during manual driving and in a lesser amount during automated driving. However, in practice, drivers also perform nondriving tasks during automated driving, of which the associated cognitive load may further reduce susceptibility to auditory information. We therefore study the effect of cognitive load during automated driving on auditory susceptibility. Method Twenty-four participants were driven in a simulated automated car. Concurrently, they performed a task with two levels of cognitive load: repeat a noun or generate a verb that expresses the use of this noun. Every noun was followed by a probe stimulus to elicit a neurophysiological response: the frontal P3 (fP3), which is a known indicator for the level of auditory susceptibility. Results The fP3 was significantly lower during automated driving with cognitive load compared with without. The difficulty level of the cognitive task (repeat or generate) showed no effect. Conclusion Engaging in other tasks during automated driving decreases auditory susceptibility as indicated by a reduced fP3. Application Nondriving task can create additional cognitive load. Our study shows that performing such tasks during automated driving reduces the susceptibility for auditory alerts. This can inform designers of semi-automated vehicles (SAE levels 3 and 4), where human intervention might be needed.

Advances in computing technology, changing policies, and slow crises are rapidly changing the way we work. Human-computer interaction (HCI) is a critical aspect of these trends, to understand how workers contend with emerging technologies and how design might support workers and their values and aspirations amidst technological change. This SIG invites HCI researchers across diverse domains to reflect on the range of approaches to future of work research, recognize connections and gaps, and consider how HCI can support workers and their wellbeing in the future.

This report documents the program and the outcomes of Dagstuhl Seminar 22102 “Computational Models of Human-Automated Vehicle Interaction”. At this Dagstuhl Seminar, we discussed how computational (cognitive) models can be used to model human-automated vehicle interaction. The seminar is motivated by developments in the field of semi-automated driving where humans and vehicles interact as teams to either both contribute to the drive (partnership) or to have safe transitions of control from vehicle to human and vice-versa. Computational (cognitive) models can be used in these situations to simulate or model human behavior and thought. Such models can be used among others to better understand human behavior, to test “what if” scenarios to guide design, or to even provide input to the vehicle about the human’s potential behavior and thoughts. The seminar was attended by experts in various fields including computer science, cognitive science, engineering, automotive UI, human-computer interaction, and human factors. They represented academia, industry, and government organizations. With the attendees, we discussed five challenges of the field during panel discussion sessions: Challenge 1: How can models inform design and governmental policy? Challenge 2: What phenomena and driving scenarios need to be captured? Challenge 3: What technical capabilities do computational models possess? Challenge 4: How can models benefit from advances in AI while avoiding pitfalls? Challenge 5: What insights are needed for and from empirical research? The attendees then split off into smaller working groups to discuss aspects of these challenges in more depth. Based on these discussions and other input from the attendees, this Dagstuhl report reports the following: an executive summary of the seminar position perspectives of all the attendees (section: “Talks”) summaries of the various working groups (section: “Working Groups”) summaries of the five panels (section: “Panel Discussions”) an overview of relevant papers (section: “Open Problems”) a research agenda with some of the most important developments and needs we identified for the field (section: “Open Problems”) All in all, we believe the seminar has shown that this field has lots of potential for development and an active community to tackle pressing issues. We can’t wait to see what results the participants of the seminar will bring to the field in the future. Seminar March 6–11, 2022 – http://www.dagstuhl.de/22102 ∗ Editor / Organizer † Editorial Assistant / Collector Except where otherwise noted, content of this report is licensed under a Creative Commons BY 4.0 International license Computational Models of Human-Automated Vehicle Interaction, Dagstuhl Reports, Vol. 12, Issue 3, pp. 15–81 Editors: Christian P. Janssen, Martin Baumann, Antti Oulasvirta, Shamsi Tamara Iqbal, and Luisa Heinrich Dagstuhl Reports Schloss Dagstuhl – Leibniz-Zentrum für Informatik, Dagstuhl Publishing, Germany 16 22102 – Computational Models of Human-Automated Vehicle Interaction 2012 ACM Subject Classification Human-centered computing → HCI design and evaluation methods; Human-centered computing → HCI theory, concepts and models; Human-centered computing → Human computer interaction (HCI); Human-centered computing; Humancentered computing → Interactive systems and tools; Human-centered computing → User models

Augmented Reality (AR) is emerging fast with a wide range of applications, including automotive AR Head-Up Displays (AR HUD). As a result, there is a growing need to understand human perception of depth in AR. Here, we discuss two user studies on depth perception, in particular on the perspective cue. The first experiment compares the perception of the perspective depth cue (1) in the physical world, (2) on a flat-screen, and (3) on an AR HUD. Our AR HUD setup provided a two-dimensional vertically oriented virtual image projected at a fixed distance. In each setting, participants were asked to estimate the size of a perspective angle. We found that the perception of angle sizes on AR HUD differs from perception in the physical world, but not from a flat-screen. The underestimation of the physical world's angle size compared to the AR HUD and screen setup might explain the egocentric depth underestimation phenomenon in virtual environments. In the second experiment, we compared perception for different graphical representations of angles that are relevant for practical applications. Graphical alterations of angles displayed on a screen resulted in more variation between individuals' angle size estimations. Furthermore, the majority of the participants tended to underestimate the observed angle size in most conditions. Our results suggest that perspective angles on a vertically oriented fixed-depth AR HUD display mimic more accurately the perception of a screen, rather than the perception of the physical 3D environment. On-screen graphical alteration does not help to improve the underestimation in the majority of cases.

This work aims to connect the Automotive User Interfaces (Auto-UI) and Conversational User Interfaces (CUI) communities through discussion of their shared view of the future of automotive conversational user interfaces. The workshop aims to encourage creative consideration of optimistic and pessimistic futures, encouraging attendees to explore the opportunities and barriers that lie ahead through a game. Considerations of the future will be mapped out in greater detail through the drafting of research agendas, by which attendees will get to know each other’s expertise and networks of resources. The two day workshop, consisting of two 90-minute sessions, will facilitate greater communication and collaboration between these communities, connecting researchers to work together to influence the futures they imagine in the workshop.

Sensory information can temporarily affect mental body representations. For example, in Virtual Reality (VR), visually swapping into a body with another sex can temporarily alter perceived gender identity. Outside of VR, real-time auditory changes to walkers’ footstep sounds can affect perceived body weight and masculinity/femininity. Here, we investigate whether altered footstep sounds also impact gender identity and relation to gender groups. In two experiments, cisgender participants (26 females, 26 males) walked with headphones which played altered versions of their own footstep sounds that sounded more typically male or female. Baseline and post-intervention measures quantified gender identity [Implicit Association Test (IAT)], relation to gender groups [Inclusion of the Other-in-the-Self (IOS)], and perceived masculinity/femininity. Results show that females felt more feminine and closer to the group of women (IOS) directly after walking with feminine sounding footsteps. Similarly, males felt more feminine after walking with feminine sounding footsteps and associated themselves relatively stronger with “female” (IAT). The findings suggest that gender identity is temporarily malleable through auditory-induced own body illusions. Furthermore, they provide evidence for a connection between body perception and an abstract representation of the Self, supporting the theory that bodily illusions affect social cognition through changes in the self-concept.