Human Perception Towards Remote Operation of Automated Vehicles

Abstract: While the development of automated vehicles (AVs) has progressed rapidly, significant challenges remain, particularly for automation levels 4 and 5 , which do not require driver intervention. Teleoperation (i.e., remote operation), where a human operator remotely controls the vehicle in critical situations, has emerged as a viable solution to these challenges. Despite this potential, the literature on user perception of the system (an essential element in understanding the needs and capabilities of teleoperation) is limited. In this context, the present study conducts a mixed-design driving simulator experiment, in which participants ride in an AV that receives assistance from remote operation during critical situations. More specifically, the type of critical event is the within-participant factor, and the remote operation level is the between-participant factor. Regarding the former, participants experience three types of critical events: human-related (a pedestrian running parallel to the road), vehicle-related (finding gaps during permissive left turns), and infrastructure-related (work zone). Regarding the latter, participants experience one of three levels of remote operation: no remote operation, remote operation with minimal information, and remote operation with extensive information. Subjective perceptions (trust, comfort, reliability, safety, security, comfort, and information sufficiency) and physiological responses (heart rate and galvanic skin response) are measured. Preliminary results show that participants who received extensive communication about teleoperation actions reported higher reliability in the vehicle's operation compared to those who received minimal information. However, trust and comfort levels did not significantly differ between the minimal and extensive information groups. Participants in both conditions expressed concerns about privacy and system security, though they generally remained open to teleoperated AV systems. The study also revealed differences in user trust based on the type of critical event, with the human-related event being deemed the most critical. These findings suggest that providing more comprehensive information during teleoperation can improve user perception of the vehicle's reliability, though other factors, such as privacy concerns and trust in the technology, persist. The study's results offer important insights for the design of teleoperated AV systems, particularly regarding communication strategies that can enhance user trust and comfort. Furthermore, additional research is needed to explore the long-term impact of teleoperation on public acceptance of AVs.