The development of automated driving systems is currently limited by narrowly defined operational design domains, restricting their deployment in real-world conditions. One promising solution to overcome limitations is teleoperation, in particular remote assistance. Remote assistance enables human intervention in ambiguous traffic situations and incorrect sensory perceptions of the vehicle where guidance or aid in decisionmaking is needed. This allows us to solve problems that exceed the vehicle's capabilities. In order to develop a functioning overall system and allow effective teleoperation, it is crucial to consider human factors for potential remote operators. Currently, there are a number of concepts about the interaction with such remote assistance systems. However, it is unclear how various interaction methods differ in terms of usability, workload, required interaction time, and which types of information are really important for remote operators. To this end, this study investigates three interaction concepts for remote assistance. Perception modification, trajectory guidance, and setting waypoints are evaluated in terms of workload (NASA-TLX), usability (UEQ-S), and preference ranking. Eye tracking data were collected to analyze the visual attention distribution of the provided information. Eighteen participants took part in a controlled lab study using an interface to provide maneuver-based solutions for six real-world driving scenarios in which the automated vehicle requested assistance. The results show that participants prefer the waypoint-based interaction method the least, while usability and perceived workload do not differ significantly between the interaction methods. All show comparable levels of performance in terms of interaction time. The remote assistance system used for this study requires improvements regarding its overall acceptance, which will be considered in the ongoing development process. This research contributes to designing an optimized remote assistance workplace, focusing on interface design, interaction efficiency, and workload reduction. The insights gained will guide further refinements in system design, ensuring that teleoperation enhances the safety and efficiency of highly automated driving in uncertain real-world conditions.
Nick et al. (Wed,) studied this question.