ABOUT

The individual mobility is becoming more and more automated, as modern assistance systems are trying to support drivers in increasingly complex driving tasks. Automated driving functions are currently developed in a technology-centric process; vehicle engineers automate whatever can be automated. The involvement of human end users in such development processes is very limited. Certainly in order to fully exploit the potential of automated driving functions - which are used voluntarily - and ultimately enhance road safety, the new technology must be truly and adequately used by drivers. However, this requires that the various user groups to gain sufficient technology acceptance and system trust.

Driving functions and assistance systems have been so far evaluated virtually as well as in simulator studies. Simulator studies often considered gender and diversity aspects very little in their sample selection process. In addition past research projects have not at all focused much on the involvement of non-professional subjects in field studies to better explore the interaction of people with (partially) automated driving functions (such as the ‘autopilot’ by Tesla Model S to name one example) in a naturalistic context. Accordingly, little knowledge has been gained so far about gender- and diversity-related differences in the interaction with (partly) automated driving functions. This bears the risk of ignoring people’s needs.

Against this background, GENDrive seeks to initiate a paradigm shift, focusing specifically on a comprehensive, scientific field study to identify gender- and diversity-related differences in requirements, system use, perceptions, acceptance and trust in the context of (partially) automated driving functions (i.e. highly automated assistance systems) to achieve a higher level of scientific rigor. GENDrive is an industrial research project integrating data science and human factors. It is focused on gaining insights through the aggregation of objective and subjective data in order to identify and quantify gender- and diversity-specific differences in system interaction in a holistic way. Above all, the interlinked analysis of differently structured data sources (e.g., data on vehicle use, driving behavior, driving styles, traffic situation, or drivers) by applying advanced methods of data science like cluster analysis and machine learning offers an immense potential for gaining insights.

Project results are

• A human factors and data-based approach to comprehensively evaluate the interaction of drivers with (semi)-automated vehicles in a naturalistic context
• A comprehensive set of qualitative and quantitative data obtained through a pre study with 10 participants and a large-scale field study with 100 participants. This data set contains data from interviews, surveys, questionnaires, thinking-aloud while driving, eye tracking, vehicle dynamics, use of assistance systems and human body sensors (ECG, EDA and acceleration sensors).
• Knowledge about drivers interacting with level 2 automation in a series vehicle that has been disseminated at several conferences. This includes also knowledge about trust and acceptance of semi-automatic driving functions. A paper presented at the 22nd International Conference on Human-Computer Interaction was awarded as best paper.
• A data analytics infrastructure and a visual analytics platform for semi-automated data analytics. Parts of the collected data can be examined with the visual analysis platform, which is accessible from this project page.

TEAM

The project is financed by

the Federal Ministry of Austria | Transport Innovation and Technology

as part of FEMtech funding

PUBLICATIONS

The project was presented at following occasions:

• Lechner G, Fellmann M, Festl A, Kaiser C, Kalayci TE, Spitzer M, Stocker A. A Lightweight Framework for Multi-device Integration and Multi-sensor Fusion to Explore Driver Distraction. In International Conference on Advanced Information Systems Engineering 2019 Jun 3 (pp. 80-95). Springer, Cham. (published paper) LINK
• Neuhuber NJ, Schnücker G, Marx C, Lechner GC, Kalayci TE, Spitzer M, Stocker A. A gender-sensitive data acquisition framework for quantification of trust and acceptance of advanced driver assistance systems. In HFES Europe chapter: Annual Meeting 2019, Oct 4, 2019, Nantes, France. (presented poster)
• Neuhuber NJ, Lechner G, Kalayci TE, Stocker A, Kubicek B. Age-related Differences in the Interaction with Advanced Driver Assistance Systems - A Field Study, In Proceedings of 22nd International Conference on Human-Computer Interaction, 19-24 July 2020, Amsterdam, Netherlands. (accepted paper, to be published)
• Lechner G. Exploring Trust in and Acceptance of Semi-Automated Passenger Vehicles. Preliminary insights obtained from a field study, Human Factor in Digital Transformation, Network at the University of Graz, 2020. (conducted talk)
• Stocker A. Exploring Trust in and Acceptance of Semi-Automated Passenger Vehicles. Results of a field study. 2nd Annual Automotive HMI and Display Forum, 13-14 November 2019, Berlin, Germany. (conducted talk)

Planned publications:

• Kalayci TE, Kalayci EG, Lechner G, Neuhuber N, Spitzer M, Stocker A, Westermeier E. Triangulated Investigation of Trust in Automated Driving: Challenges and Solution Approaches for Data Integration. Journal of Industrial Information Integration. (planned journal paper)
• Stocker A and Neuhuber N. Driver Interaction tracking in a real car environment. World Usability Congress, 21-22 Oct 2020, Graz. Austria (planned talk)