Mediating between Human Driver and Automation through Human-Machine Interface for Highly Automated Vehicles

Abstract

This master thesis report gathers the research and design activities executed in 100 working days, in order to investigate the future context of autonomous driving. More specifically, the communication between the human driver and automation during the automated driving modes (Long Out of the Loop and Standby modes) has been investigated to answer the knowledge gaps of mode confusion, transparency, and information load. The project was supervised by SWOV, the Dutch Institute for Road Safety Research.
The project approach refers to a double-diamond structure which includes firstly, context research as background knowledge about autonomous driving and HMI. Secondly, an extensive user research involved experts, drivers, and human factors as the core of the study. Thirdly, an Experience Journey Map was developed to set bases for the design definition and Functional requirements, and HMI qualities. The design goal for the project has been based on the research results, aiming to unobtrusively inform drivers about automation status, driving mode and their responsibilities during autonomous driving mode through the HMI, while the driver is performing non-driving related activities. The HMI concepts have been developed according to iterative conceptualization and evaluation phases to conduct experiments with participants evaluating lo-fi prototypes about singular interactions and design elements, collecting conclusions, and highlighting strengths for each concept. Those helped to put together a final concept that includes the best elements of previous phases.
The final design proposal is a holistic experience that incorporates together different HMI options such as ambient lighting, central display, variable windshield dimming, the windshield visual effects and haptic feedback in order to create a complete experience to test with a sample of participants.
The final user tests have been conducted digitally with participants who evaluated the interface of automated driving modes while participants were asked to focus on secondary tasks. The tests showed that the ambient lighting (together with adjustable windshield dimming) of the vehicle certainly helps to set the driver’s mindset and support secondary tasks. The messages on the windshield and 3D glow effects obtained good results in terms of transparency. Drivers, indeed, understood more information about road situations and automation behavior. This encouraging results open horizons to new applications about ambience and visual information. Globally, the new information presented to participants was perceived as pleasant, informative and unobtrusive. Participants accepted the new HMI options with confidence and enthusiasm, although, some aspects of the central display were not fully understood in terms of usability.

In conclusion, the design research conducted in the first part of the project brought a fresh view of the user’s perspective. The insights from the human-centered design approach are extremely useful to investigate the communication to establish between humans and automation. This master thesis also contributed to bring some initial but promising results in the new direction of HMI design for Mediator project.