Dissertation
Exploration of smart infrastructure for drivers of autonomous vehicles: design space, out-of-view visualization and explanations
Vorschaubild nicht verfügbar
Volltext URI
Dokumententyp
Text/Dissertation
Zusatzinformation
Datum
2020
Autor:innen
Zeitschriftentitel
ISSN der Zeitschrift
Bandtitel
Quelle
Verlag
Ludwig-Maximilians-Universität München, Fakultät für Mathematik, Informatik und Statistik
Zusammenfassung
The connection between vehicles and infrastructure is an integral part of providing autonomous vehicles information about the environment.
Autonomous vehicles need to be safe and users need to trust their driving decision. When smart infrastructure information is integrated into the vehicle, the driver needs to be informed in an understandable manner what the smart infrastructure detected.
Nevertheless, interactions that benefit from smart infrastructure have not been the focus of research, leading to knowledge gaps in the integration of smart infrastructure information in the vehicle. For example, it is unclear, how the information from two complex systems can be presented, and if decisions are made, how these can be explained.
Enriching the data of vehicles with information from the infrastructure opens unexplored opportunities.
Smart infrastructure provides vehicles with information to predict traffic flow and traffic events.
Additionally, it has information about traffic events in several kilometers distance and thus enables a look ahead on a traffic situation, which is not in the immediate view of drivers.
We argue that this smart infrastructure information can be used to enhance the driving experience. To achieve this, we explore designing novel interactions, providing warnings and visualizations about information that is out of the view of the driver, and offering explanations for the cause of changed driving behavior of the vehicle.
This thesis focuses on exploring the possibilities of smart infrastructure information with a focus on the highway.
The first part establishes a design space for 3D in-car augmented reality applications that profit from smart infrastructure information. Through the input of two focus groups and a literature review, use cases are investigated that can be introduced in the vehicle's interaction interface which, among others, rely on environment information. From those, a design space that can be used to design novel in-car applications is derived.
The second part explores out-of-view visualizations before and during take over requests to increase situation awareness. With three studies, different visualizations for out-of-view information are implemented in 2D, stereoscopic 3D, and augmented reality. Our results show that visualizations improve the situation awareness about critical events in larger distances during take over request situations.
In the third part, explanations are designed for situations in which the vehicle drives unexpectedly due to unknown reasons. Since smart infrastructure could provide connected vehicles with out-of-view or cloud information, the driving maneuver of the vehicle might remain unclear to the driver. Therefore, we explore the needs of drivers in those situations and derive design recommendations for an interface which displays the cause for the unexpected driving behavior.
This thesis answers questions about the integration of environment information in vehicles'.
Three important aspects are explored, which are essential to consider when implementing use cases with smart infrastructure in mind. It enables to design novel interactions, provides insights on how out-of-view visualizations can improve the drivers' situation awareness and explores unexpected driving situations and the design of explanations for them.
Overall, we have shown how infrastructure and connected vehicle information can be introduced in vehicles' user interface and how new technology such as augmented reality glasses can be used to improve the driver's perception of the environment.