The influence of AR information on pedestrians' situational awareness when using autonomous mobility services

Abstract

Driven by increasing urbanisation and the growing demand for sustainable and accessible transport solutions, future mobility systems are likely to be complemented by Shared Automated Mobility-onDemand (SAMoD) services. These systems provide flexible, infrastructure-independent travel options by assigning dynamic virtual stops (vStops) where users walk a short distance to board a driverless vehicle. While vStops offer new opportunities for efficient and scalable mobility, they also pose challenges particularly for pedestrians. These include identifying the vehicle and pick-up point, as well as safely orienting oneself in complex and unknown traffic situations. To support interaction with other road users, wearable augmented reality - human-machine interfaces (AR-HMIs) have the potential to provide intuitive, context-sensitive guidance. However, despite the growing interest in AR technologies, user-centered AR-HMIs for SAMoD scenarios remain largely unexplored. This study addresses this gap by investigating how the amount and type of visual AR cues in wearable AR-HMIs affect pedestrians' situational awareness (SA), user experience (UX), and acceptance in pickup scenarios. The primary objective is to determine the optimal amount of information to provide to pedestrians without overwhelming them with excessive cues. A video-based online study was conducted with 187 participants (N = 106 aged 18-35 years, N = 81 aged 65 years and older) who evaluated different AR-HMI designs embedded in simulated traffic environments. Each participant was presented with different AR-HMI designs in two traffic complexity levels (low vs. high) to rate in terms of SA, UX and acceptance. The results show that simple AR cues significantly improve SA, especially for younger users. In contrast, more complex, information-rich AR-HMI designs lead to a significant decrease in usability and acceptance. Apart from more complex traffic situations, older users show a lower sensitivity to the variations in AR cues without any statistically significant differences. Younger participants prefer simple AR-HMI designs with fewer cues. Older users, in contrast, show no clear preference for a specific AR-HMI design. These findings highlight the importance of designing age- and context-sensitive AR-HMIs and suggest that minimal, targeted information may be more effective than complex, multi-layered visualisations in dynamic urban environments. This study contributes to the emerging field of wearable AR-HMIs in SAMoD contexts by providing initial insights into designing ARHMIs that consider diverse user needs and the presence of multiple road users.