Do We Still Need UI for Self-Driving Cars?

If you haven't got a driving license, think twice. Car vendors are rushing to get self-driving cars on the road. Once they can substitute the "autos", chauffeur as a profession will not be needed so much any more. Instead of learning driving, please consider alternatives, for example, spending your time in getting self-driving experience more comfortable.

The research area of Dr Pierre Akiki is engineering model-driven self-adaptive user interfaces for enterprise software systems, at the crux of software engineering and human-computer interaction. In less than three years, he has published top journal papers in ACM Computing Surveys, IEEE Transactions on Software Engineering, as well as top conferences in HCI and Software Engineering. Don't worry if you have no time or interests in reading the details of his publications, in a nutshell, the entire PhD study can be summarised into one research question: how to model self-adaptive user interfaces and use the model to simplify human computer interactions while adapting the interface to best suite different usage contexts?

Software systems used to compete for all-inclusive bloat features. Self-adaptive user interfaces are the contrary. By leaving out unnecessary options automatically, and bringing them back only when needed, software systems become so much more usable. Our engineering practices have demonstrated the improved usability for large-scale software systems for the enterprises with very little execution overhead.

Let's imagine the user interface of a self-driving car. It shall be much simpler than a traditional car. Wait a second, isn't it fully automated so there is no more need for human interventions, notwithstanding the human interfaces? Wrong! Once you are on the driving seat, you would not be idle! As long as you demand the car to do something, something functional to satisfy you, you would need such an interface!

However, you used to concentrate once sitting on the driving seat, otherwise you will be responsible for any incident caused by your distractions. With self-driving software in charge of safety, you can do more: making phone calls, checking emails, surfing internet, playing games, writing blogs, and so on. Applying Pierre's principle, "less is more", how do we simplify these tasks, then? In other words, when life is getting more and more complex, how can you trust a user interface to take over mundane tasks in managing the life complexity? So that you have a bit more time on thinking, focusing on the flow that requiring your focus.

Sitting inside self-driving cars, we can ask a number of research questions. For example, are there different contexts-of-use that require the same UI to adapt? What kind of bloatness exists in the UIs of self-driving cars, and according to what contexts-of-use should we reduce this bloat to achieve simplicity? Could we use the profile of each passenger in the cloud? Could there be a more natural mode of interaction, e.g., using a voice chat modality, rather than touch screen? This poses the question of whether UI adaptation applies, and if so, what kind of adaption would be required? Probably the car could pose the questions in a different way depending on the age of the passenger as well...

This list of questions is not exhaustive. What we need to think and act now is to advance the user interfaces research so that it is ahead of the era when self-driving becomes the default option. Simplifying your life journey, it may help you achieve more.

References

Chinese translation
Email: y.yu@open.ac.uk Office: +44 (0) 1908 6 55562