Autonomous Driving - Analysing the Impact of Resilience Engineering Features in Smart Car Interfaces

Abstract

The communication between driver and vehicle plays an important role in the development of car automation. Being able to perform non-driving related tasks in uncritical situations is one of the biggest advantages of autonomous driving in smart cars. Yet, critical situations often require the driver to take over control again, thereby making the design of takeover maneuvers a highly relevant task. Our study thus investigated the impact of basic vs. enhanced HMIs considering the principles of resilience engineering on supporting drivers in taking control over the vehicle. We investigated driving on the highway in two critical scenarios: a roadworks and a potential collision scenario. In a VR study in a driving simulator 45 participants tested three HMIs: basic, visual, and visual + speech-based. Even though the results indicate high usability scores and a positive user experience for all HMIs, no significant differences regarding takeover time could be measured between the HMIs. Reasons include the visual differences in the interface design with regards to resilience engineering being minor and the sample size being too small to detect small effects. Implications for future research and the design of takeover maneuvers are discussed.