Accessible Automated Automotive Workshop Series – Accessibility in Mobility. In conjunction with MuC 2025

Terminology

Automated Vehicles (AVs): Vehicles that provide some level of automation, thereby reducing the effort of the human driver for the primary driving task. In this workshop, we focus on AVs that do not need an attentive driver at all times, thus SAE Levels 3 and above.

Vulnerable Road Users (VRUs): Non-motorized road users, such as pedestrians and cyclists, motorcyclists, and persons with disabilities, reduced mobility, or reduced orientation [Connell1997]. This workshop specifically focuses on especially vulnerable road users eVRUs [Hollaender2021] like children, older adults or people with impairments.

External Human-Machine Interfaces (eHMIs): Interfaces such as LED bands, screens, or loudspeakers that an automated vehicle is equipped with to communicate with people outside the vehicle.

Background

In the near future, AVs will alter traffic for all road users: removing human drivers from the car will diminish human failure [17], allow for non-driving related tasks within the vehicle [31], and change communication needs for other road users such as pedestrians or cyclists [21].

Automated Vehicles

With ongoing deployment of AVs [30], traffic is anticipated to become increasingly automated. Beyond the technical challenges of automation [24], effective interaction with other road users — such as pedestrians or cyclists [36] — must be both efficient [25] and trusted [38]. These factors are especially critical for PSN to engage in daily life activities [2,35], including shared spaces. Reduced mobility access can limit social opportunities, healthcare availability, shopping, or even employment [3,33]. Consequently, improved automation is viewed as a key enabler for enhancing PSN participation in everyday life.

Communication Between Road Users

Communication between road users is vital for preventing conflicts and ensuring traffic safety. By employing signals, hand gestures, and eye contact, along with implicit cues such as vehicle behavior (e.g., slowing down), individuals convey intentions that help others anticipate and adapt [26]. As driving automation replaces human drivers, these communication cues may be diminished. To address this gap, researchers propose external human-machine interfaces (eHMIs) [15].

Previous studies have investigated explicit design spaces [11], integrated environmental contexts such as sidewalks [9], identified scalability limitations [14], analyzed use cases like construction sites [10], conducted systematic comparisons [28], and introduced anthropomorphic features [5,7]. However, existing research still underrepresents inclusive communication practices for AVs and vulnerable road users (VRUs) [12,13,19].

Focus on Vulnerable Road Users

VRUs are defined by the World Health Organization as pedestrians, cyclists, and motorcyclists [39] and are generally characterized by the absence of a protective external shield [37]. However, Holländer et al. [21] advocate a more granular understanding of VRUs in traffic and Human-Machine Interaction contexts. Motorization is a primary distinguisher (e.g., motorcyclist or personal conveyance vs. pedestrian or cyclist), and each group may be “especially vulnerable” due to factors such as age or disability [21]. This workshop focuses on these especially vulnerable segments to promote more accessible (future) traffic systems.

As populations age and survival rates from accidents and diseases increase, accessibility for diverse abilities becomes even more critical. Demographic shifts already limit the availability of medical and care personnel, making it difficult to provide significant assistance beyond essential support. Thus, inclusive automated traffic is not only ethically imperative but also vital for maintaining the independence of people with disabilities who depend on mobility for crucial tasks such as shopping or reaching medical appointments.

Design Processes to Support People with Special Needs

To foster an added value for PSN, processes must be centered around their specific needs [29]. While the Universal Design approach [18] aims to design products accessible for all people, Inclusive Design [6] tries to target barriers and implement countermeasures. While the outcome of both approaches may be similar, including PSN in research is already a significant challenge. For instance, making surveys accessible for both visually [23] and intellectually [20] impaired people increases the effort and expenses for studies.

In recent years, AI was increasingly used to support design processes and the inclusivity of research methods [16], but primarily in education [1,32,34]. Human-in-the-loop UI optimization (e.g., [4,8,22]) can adapt designs to each user’s preferences, needs, and abilities based on their iterative feedback. However, it remains unclear how PSN can provide input, their desired level of involvement, and how they understand a design optimization process with diverse information needs. Moreover, in a multimodal driving environment, where multiple road users might interact with the same UI (e.g., eHMI), it remains challenging to determine whose needs a design (optimization) process should prioritize.