AI Lab Trust

Human AI systems: Interaction design for interiors of autonomous driving systems in public transport

This project is dedicated to the topic of trust in human-AI interaction, which is necessary for the acceptance of autonomous transport systems in public transport. The aim is to develop an interdisciplinary methodology for the co-design of AI systems.

Therefore, design research and systems engineering were brought together in order to determine which factors influence trust in autonomous vehicles. As an experimental setup, a virtual reality simulation was developed in which an autonomous vehicle is driving in mixed traffic and is forced to brake fully in a risky situation. In order to investigate the subjects’ perception of the environment and their reaction, biometric data were collected and personal attitudes measured. Four different times of an alerting were tested using specific design principles. By doing so, it was possible to determine that the timing of the alerting influences situational awareness and thus overall confidence. Based on such tests, new conclusions can be drawn about the correlations between design parameters and scenarios of a potential breach of trust and can be integrated into the development of AI systems.

The project (HA project no. 1014/21-16) was funded by the state of Hesse and HOLM funding as part of the “Innovations in Logistics and Mobility” measure of the Hessian Ministry of Economics, Energy, Transport and Housing.

Type of project • Forschungsprojekt
Lead • Prof. Peter Eckart, Prof. Dr. Kai Vöckler
Team • Dipl.-Des. Ken Rodenwaldt, Dipl.-Des. Armin Arndt, Dipl.-Des. Hannah Weirich, Leonard Neunzerling
Status • completed
Duration • 05/2021 – 04/2022
Cooperation • Prof. Dr. Visvanathan Ramesh, Dr. Iuliia Pliusch, Martin Klingebiel, Goethe-Universität Frankfurt
Funding • State of Hessen/House of Logistics and Mobility (HOLM), Hessian Ministry of Economics, Energy, Transport and Housing

More...

Publications about the project

No data was found

Cooperation

Funding

Consent Management Platform by Real Cookie Banner Skip to content