Figure 1: The 233 countries and territories with dashcam footage in CROWD dataset [2].The colouration is based on the logarithm of the total recorded time per country or territory, calculated as log 𝑒 (1 + time in seconds), to reduce the skew from outliers such as the United States (with 707.76 hours available).The black dots show the 2,495 cities in the dataset.The labels under images show the corresponding YouTube video ID.The frame on the bottom left shows an example of object detection using YOLOv11x with identified objects such as pedestrians, vehicles, and traffic signs; in this image, the labels 'id' refer to the unique ID of the detected object with the type mentioned later and end with the confidence of detection of the object.

Drones will soon deliver packages to recipients in public spaces, where drones may encounter difficulties identifying safe drop-off locations. Such uncertainties can reduce trust and raise safety concerns. This augmented reality study investigates how recipients perceive being asked to guide the drone in uncertain situations using hand gestures, and to what extent they feel comfortable with different levels of involvement. Results show that participants preferred a basic level of involvement, which received higher trust and usability scores than either no or high involvement. We recommend involving recipients in the final stage of delivery to not only support drone operations but also improve recipient trust and clarity in uncertain conditions.

Drones are likely to enter social spaces in the foreseeable future. Novel Human-Drone Interactions (HDI) will foster beyond typical drone-operator interaction, posing new human factors challenges. However, the specific focus areas for HDI research remain unclear. This study conducts 11 expert interviews to identify potential use cases and human factors challenges for HDI in public spaces. Initial drone use cases include emergency response and delivery scenarios, where the general public may interact as recipients and bystanders, each posing unique challenges. Uncertainty, stemming from a lack of awareness, emerges as a significant human factors concern, impacting perceived risk. Addressing this uncertainty, especially in recipients, may involve refining drone behaviour, physical appearance, and interface design. The challenges identified in this study lay the groundwork for future HDI research in public spaces.

The use of drones is expected to increase for delivering groceries or medical equipment to individuals. Understanding how people perceive drone behavior, specifically in terms of approach trajectories and delivery methods, and identifying factors that induce feelings of uncertainty is crucial for perceived safety and trust. This virtual reality experiment investigated the impact of drone approach trajectories and delivery methods on feelings of uncertainty. Forty-five participants observed a drone approaching in an orthogonal or a curved path and either, delivering packages by landing or using a cable while hovering above eye level. We found that participants felt uncertain and unsafe, especially when looking up at drones approaching with orthogonal paths. Curved paths led to lower feelings of uncertainty, with comments such as being more natural, trustful, and safe. Feelings of uncertainty arose while landing on the ground due to altitude changes and potential collision concerns. Using a cable instead of actually landing for delivery reduced feelings of uncertainty and increased trust. The study recommends drones avoid hovering near humans, especially after landing. Furthermore, the study suggests exploring design solutions, including design aesthetics and human–machine interfaces, that clearly convey drone intentions to help reduce feelings of uncertainty.

The advancement of automated driving technologies is fundamentally transforming the relationship between humans and vehicles, shifting from direct control to a more collaborative dynamic. Consequently, the design of in-vehicle Human-Machine Interfaces (iHMIs) is becoming increasingly intricate, focusing on aspects beyond mechanics and ergonomics towards enriched interaction and enhanced user experience. This shift has prompted research efforts to explore and advance iHMI concepts. Despite the iterative nature of design and its role in knowledge creation, our high-level understanding of the design processes utilised in iHMI development remains limited. To provide a comprehensive overview, this paper presents a scoping review of 324 papers (2013—2023) focused on the design underpinnings of iHMI development. Our review presents a categorisation of study goals and a detailed classification of five key stages within the design process. Based on these analyses, we discuss the influence of design and identify potential avenues for future research on iHMIs.

Drones will likely enter public spaces soon for deliveries, interacting with humans as recipients or bystanders. Understanding their requirements and factors contributing to uncertainty is crucial for safer Human-Drone Interaction design. Twelve participants were interviewed and engaged in focus groups. Preliminary results indicate the need for clear communication of drone intentions through design and interfaces, guiding future research and design concepts to be tested with user studies for delivery drones in public spaces.

Current research on external Human-Machine Interfaces (eHMIs) in facilitating interactions between automated vehicles (AVs) and pedestrians have largely focused on one-to-one encounters. In order for eHMIs to be viable in reality, they need to be scalable, i.e., facilitate interaction with more than one pedestrian with clarity and unambiguity. We conducted a virtual-reality-based empirical study to evaluate four eHMI designs with two pedestrians. Results show that even in this minimum criteria of scalability, traditional eHMI designs struggle to communicate effectively whom the AV intends to yield to. Road-projection-based eHMIs show promise in clarifying the specific yielding intention of an AV, although it may still not be an ideal solution. The findings point towards the need to consider the element of scalability early in the design process, and potentially the need to reconsider the current paradigm of eHMI design.

In Level 3 automated vehicles, drivers must take back control when prompted by a Take Over Request (TOR). However, there is currently no consensus on the safest way to achieve this. Research has shown that participants interact faster with an avatar when this “glows” in synchrony with participant physiology (heartbeat). We hypothesized that a similar form of synchronization might allow drivers to react faster to a TOR. Using a driving simulator, we studied driver responses to a TOR when permanently visible ambient lighting was synchronized with participants’ breathing. Experimental participants responded to the TOR faster than controls. There were no significant effects on self-reported trust or physiological arousal, and none of the participants reported that they were aware of the manipulation. These findings suggest that new ways of keeping the driver unconsciously “connected” to the vehicle could facilitate faster, and potentially safer, transfers of control.

External human-machine interfaces (eHMIs) support automated vehicles (AVs) in interacting with vulnerable road users such as pedestrians. While related work investigated various eHMIs concepts, these concepts communicate their message in one go at a single point in time. There are no empirical insights yet whether distance-dependent multi-step information that provides additional context as the vehicle approaches a pedestrian can increase the user experience. We conducted a video-based study (N = 24) with an eHMI concept that offers pedestrians information about the vehicle’s intent without providing any further context information, and compared it with two novel eHMI concepts that provide additional information when approaching the pedestrian. Results show that additional distance-based information on eHMIs for yielding vehicles enhances pedestrians’ comprehension of the vehicle’s intention and increases their willingness to cross. This insight posits the importance of distance-dependent information in the development of eHMIs to enhance the usability, acceptance, and safety of AVs.

In this paper, we report user preferences regarding color and animation patterns to support the interaction between Automated Vehicles (AVs) and pedestrians through an external Human-Machine-Interface (eHMI). Existing concepts of eHMI differ -- among other things -- in their use of colors or animations to express an AV's yielding intention. In the absence of empirical research, there is a knowledge gap regarding which color and animation leads to highest usability and preferences in traffic negotiation situations. We conducted an online survey (N=400) to investigate the comprehensibility of a light band eHMI with a combination of 5 color and 3 animation patterns for a yielding AV. Results show that cyan is considered a neutral color for communicating a yielding intention. Additionally, a uniformly flashing or pulsing animation is preferred compared to any pattern that animates sideways. These insights can contribute in the future design and standardization of eHMIs.

In road-crossing situations involving negotiation with approaching vehicles, pedestrians need to take into account the behavior of the car before making a decision. To investigate the kind of information about the car that pedestrians seek, and the places where do they look for it, we conducted an eye-tracking study with 26 participants and analyzed the fixation behavior when interacting with a manually-driven vehicle that approached while slowing and displaying a yielding behavior. Results show that a clear pattern of gaze behavior exists for pedestrians in looking at a vehicle during road-crossing situations as a function of the vehicle's distance. When the car is far away, pedestrians look at the environment or the road space ahead of the car. With the approach, the gaze gradually shifts to the windshield of the car. We conclude by discussing the implications of this insight in the user-centered-design of optimal external Human-Machine-Interfaces for automated vehicles.