The Predictive Corridor: A Virtual Augmented Driving Assistance System for Teleoperated Autonomous Vehicles

Abstract

Autonomous vehicles offer a driverless future, however, despite the rapid progress in ubiquitous technologies, human situational assessment continues to be required. For example, upon recognizing an obstacle on the road a request might be routed to a teleoperator, who can assess and manage the situation with the help of a dedicated workspace. A common solution to this problem is direct remote steering. Thereby a key problem in teleoperation is the time latency and low remote situational awareness. To solve this issue we present the Predictive Corridor (PC), a virtual augmented driving assistance system for teleoperated autonomous vehicles. In a user study (N =32), we evaluated the PC by employing three measures: performance, subjective and physiological measures. The results demonstrate that driving with the PC is less cognitively demanding, improves operational performance, and nonetheless can visually compensate for the effect of the time delay between the teleoperator and the vehicle. This technology, therefore, is promising for being applied in future teleoperation applications.