2016-2020 (My Dissertation)
The individual and societal benefits of driving automation can only unfold if the underlying technology is established on the market. As user acceptance is dependent on users’ experience with a technology, i.e. user experience (UX), novel user interfaces (UIs) need to be developed to balance drawbacks of individual automation levels (SAE J3016). Therefore, the predominant innovation- and technology-centered perspective has to be supplemented by a user-centered approach.
I used a three-fold approach to study driving automation UX and develop an optimized development approach for driving automation UIs. Therefore, the following research questions (RQs), that build upon each other, were put forward:
RQ1: How is UX in the context of driving automation methodologically addressed in practice?
RQ2: How can UX theory and the insights from UX practice be combined to optimize the development process of driving automation systems?
RQ3: How must user interfaces be designed to positively affect UX of driving and while being driven?
What Could I Learn from UX Practice (RQ1)?
After reviewing UX theory and related work in the field, which showed that a satisfactory systematic approach to study driving automation UX is still missing,
UX practice of academia and industry (RQ1) was analyzed in more detail.
We conducted two literature reviews, one for UX research in general and another for driving automation in general, and a literature study with UX practitioners from industry working on driving automation UX. Results indicate that there is still no consistent understanding across departments and disciplines what UX means. This leads to technology- instead of need-centered processes in which organizational issues still prevent a real iterative design process. As it is unclear which aspects need to be studied to access UX, mostly traditional methods like interviews and self-defined questionnaires are conducted, without a clear strategy on how to apply triangulation in a meaningful way to get a more reliable, holistic and well-motivated understanding of UX. The temporality of UX is mainly ignored and quantitative results are favored.
How Could UI Development be Optimized (RQ2)?
To improve UX practice (RQ2) my thesis presented the “DAUX Framework” which is part of a need-centered development approach. This was inspired by UX theory and related studies. The focus of this approach is on fulfilling users’ psychological needs whose prioritization differs regarding context and situation. The framework offers guidelines how to a) identify relevant needs for hypothesis/ UI concept development and b) evaluate UX by triangulating behavioral, product-, and experience-oriented methods. Therefore, sequential triangulation, by using qualitative approaches to reveal which needs are relevant in a certain context and level of automation, and evaluation of a developed user interface example by mixing behavioral-, product-, and experience-oriented data as concurrent triangulation strategy are recommended.
Thereby, the “DAUX Framework” gives a guideline by unbundling and visualizing related components of UX, which automatically implies a certain method selection to study UX from different perspectives.
How to Design Driving Automation UIs (RQ3)?
The need-centered process was then applied in three case studies to develop example user interfaces which improve driving automation UX (RQ3). Each case study regards a different level of automation (SAE L2, L3, and SAE L4/5), however, all investigated the context of highway diving. The framework helped to substantiate hypotheses and facilitated discussions in the interdisciplinary team about the right method selection to study UX relevant constructs for evaluation. Results of the presented case studies are important findings for driving automation UI development.
Regarding SAE L2 (see “In UX we Trust”), overtrust/ overreliance has already led to fatal accidents. Thus partial driving automation may not be safely possible without making system performance accessible to drivers, e.g., by reliability displays. The results of our study give insights in how the stream of experiences combining performance, usability, and aesthetics of different vehicle subsystems correlate and influence each other. This emphasizes the importance of a joint contemplation of driving automation UX by an interdisciplinary team of UX, trust, and safety researchers.
Further, the safety problem of SAE L3 driving was highlighted by various studies. A lot of effort is invested to reduce take-over times. However, combining methods from safety but also emotional research shows the difficulty also from an experience perspective. Results (see project ATHENA) show that due to users’ needs for autonomy, competence, and security, the mere possibility of a take-over-request at any time negatively impacts the whole journey experience.
At high and full driving automation (SAE L4/5, see project “The Driving Hotzenplotz”), users worry about their needs of competence, autonomy, and the meaning of driving interactions, e.g., accelerating. Although engaging in non-driving related tasks might balance these problems, there will still be users who appreciate the joy of driving. Hence, optional control should always be offered. The presented interface for optional cooperative control is no real solution that should be built like our prototype, however, it is an example of how an increased user autonomy can enhance the joy of driving, thus, user experience. Hence, optional control should always be offered.
Conclusion
Driving automation comes with many promises like increased road safety, improved traffic flow, mobility for new target groups, and more leisure time. But there is still skepticism in the society which need to be carefully regarded because it can prevent a holistic establishment of driving automation systems on the market.
Especially in the crisis of the automotive industry, traditional processes have to be revised and the focus shifted from money-making to the fulfillment of human needs. Thereby, users are not professionals who have to deal with human-related issues of automation. As users spend their private time and money, their expectations on having positive experiences in the car have to be met. This is important to regain trust and to bind customers for the future. Hence, all promises which are given by the concept of driving automation require advanced technology, but in addition intelligent and well-designed UI concepts to support the user during the drive. A focus on users’ psychological needs helps to create positive experiences in the different levels of automation.
While still many questions regarding driving automation are unanswered and circumstances unknown, the “DAUX Framework” embedded in a need-centered development approach provides a guideline for driving automation research and development. This will iteratively improve driving automation user experiences for the future and bring mobility to the next higher level of innovation.
Dr. Anna-Katharina Frison 