Drowsy driving imposes a high safety risk. Current systems often use driving behavior parameters for driver drowsiness detection. The continuous driving automation reduces the availability of these parameters, therefore reducing the scope of such methods. Especially, techniques that include physiological measurements seem to be a promising alternative. However, in a dynamic environment such as driving, only non- or minimal intrusive methods are accepted, and vibrations from the roadbed could lead to degraded sensor technology. This work contributes to driver drowsiness detection with a machine learning approach applied solely to physiological data collected from a non-intrusive retrofittable system in the form of a wrist-worn wearable sensor. To check accuracy and feasibility, results are compared with reference data from a medical-grade ECG device. A user study with 30 participants in a high-fidelity driving simulator was conducted. Several machine learning algorithms for binary classification were applied in user-dependent and independent tests. Results provide evidence that the non-intrusive setting achieves a similar accuracy as compared to the medical-grade device, and high accuracies (>92%) could be achieved, especially in a user-dependent scenario. The proposed approach offers new possibilities for human–machine interaction in a car and especially for driver state monitoring in the field of automated driving.

During the last decade, research has brought forth a large amount of studies that investigated driving automation from a human factor perspective. Due to the multitude of possibilities for the study design with regard to the investigated constructs, data collection methods, and evaluated parameters, at present, the pool of findings is heterogeneous and nontransparent. This literature review applied a structured approach, where five reviewers investigated n = 161 scientific papers of relevant journals and conferences focusing on driving automation between 2010 and 2018. The aim was to present an overview of the status quo of existing methodological approaches and investigated constructs to help scientists in conducting research with established methods and advanced study setups. Results show that most studies focused on safety aspects, followed by trust and acceptance, which were mainly collected through self-report measures. Driving/Take-Over performance also marked a significant portion of the published papers; however, a wide range of different parameters were investigated by researchers. Based on our insights, we propose a set of recommendations for future studies. Amongst others, this includes validation of existing results on real roads, studying long-term effects on trust and acceptance (and of course other constructs), or triangulation of self-reported and behavioral data. We furthermore emphasize the need to establish a standardized set of parameters for recurring use cases to increase comparability. To assure a holistic contemplation of automated driving, we moreover encourage researchers to investigate other constructs that go beyond safety.

Numerous statistics show that cyclists are often involved in road traffic accidents, often with serious outcomes. One potential hazard of cycling, especially in cities, is “dooring”—passing parked vehicles that still have occupants inside. These occupants could open the vehicle door unexpectedly in the cyclist’s path—requiring a quick evasive response by the cyclist to avoid a collision. Dooring can be very poorly anticipated; as a possible solution, we propose in this work a system that notifies the cyclist of opening doors based on a networked intelligent transportation infrastructure. In a user study with a bicycle simulator (N = 24), we examined the effects of three user interface designs compared to a baseline (no notifications) on cycling behavior (speed and lateral position), perceived safety, and ease of use. Awareness messages (either visual message, visual message + auditory icon, or visual + voice message) were displayed on a smart bicycle helmet at different times before passing a parked, still-occupied vehicle. Our participants found the notifications of potential hazards very easy to understand and appealing and felt that the alerts could help them navigate traffic more safely. Those concepts that (additionally) used auditory icons or voice messages were preferred. In addition, the lateral distance increased significantly when a potentially opening door was indicated. In these situations, cyclists were able to safely pass the parked vehicle without braking. In summary, we are convinced that notification systems, such as the one presented here, are an important component for increasing road safety, especially for vulnerable road users.

Scenario-based proving ground testing has been established as the standard test environment for the homologation of driver assistance systems. These test procedures are described, for example, in the New Car Assessment Programme test catalog, which only considers a limited number of possible scenarios. Furthermore, pedestrian targets are, among others, used to test assistance systems for protection against vulnerable road users (VRUs). However, these targets also have the disadvantage that they only represent humans in a simplified way and cannot depict realistic movement patterns. On the other hand, simulation has been established as an efficient means of realizing scenarios in a reproducible, time- and cost-saving manner. However, simulation has a major disadvantage: it can only represent reality to a limited extent, e.g., in the influence of human factors in response to hazardous situations. To bridge the gap between simulation and real-world testing, we introduce MiRE, a mixed reality environment for testing automated driving functions. In MiRE, real entities meet in a virtual environment and take advantage of real and simulated test procedures. This paper presents MiRE, a proof of concept based on a modular architecture. The goal is to use it to evaluate the feasibility of testing in mixed reality – by comparing it to proving ground testing. Preliminary results show that MiRE is a suitable environment for testing complex scenarios where the real vehicle and VRUs can safely interact in safety-critical situations and produce realistic results.

Human-centered XAI (HCXAI) advocates that algorithmic transparency alone is not sufficient for making AI explainable. Explainability of AI is more than just “opening” the black box — who opens it matters just as much, if not more, as the ways of opening it. In the era of Large Language Models (LLMs), is “opening the black box” still a realistic goal for XAI? In this fourth CHI workshop on Human-centered XAI (HCXAI), we build on the maturation through the previous three installments to craft the coming-of-age story of HCXAI in the era of Large Language Models (LLMs). We aim towards actionable interventions that recognize both affordances and pitfalls of XAI. The goal of the fourth installment is to question how XAI assumptions fare in the era of LLMs and examine how human-centered perspectives can be operationalized at the conceptual, methodological, and technical levels. Encouraging holistic (historical, sociological, and technical) approaches, we emphasize “operationalizing.” We seek actionable analysis frameworks, concrete design guidelines, transferable evaluation methods, and principles for accountability.