Design of a High-Level Guidance User Interface for Teleoperation of Autonomous Vehicles

Autonomous vehicles (AVs) are a disruptive mode of transportation that is rapidly advancing. However, it is widely acknowledged in industry and academia that AVs may not be capable of handling every traffic situation independently, necessitating remote human intervention. Existing teleoperation methods face significant challenges, highlighting the need for innovative remote operation approaches. One such approach is tele-assistance, where remote operators (ROs) offer high-level guidance while delegating low-level controls to automation. Our research focuses on designing a tele-assistance interface. By interviewing 14 AV teleoperation experts and conducting an elicitation study with 17 experienced teleoperators, we identify road scenarios requiring remote human assistance. We then devise a set of discrete high-level commands that enable the resolution of these scenarios without manually controlling AVs. Finally, we integrate these findings into the design of a novel user interface for teleoperating autonomous vehicles.