The SDV needs a new concept for usability. Here is why.

I have to admit. I am guilty. Like most tech people, I think of technical innovation and get excited and immediately want to try it out. Many of us engineers and computer scientists live in a tech-savvy bubble - the majority of people do not belong to this bubble.

I know people that have ADAS systems in their vehicle and never use them because they fear the distraction of operating them while driving - kind of ironic when you think of what those systems are made for. It becomes obvious that we have a usability problem.

Now imagine, we did have real software-defined vehicles. The ones, where a user can decide to install new software. Where there is new functionality available - potentially even developed by a third party. Who would be the users and how would you use that functionality - and I am talking about mass market adoption and not the tech-savvy nerd.

The over-used smartphone comparison

As old as it gets, I still want to use the good old smartphone analogy. Back in the days before the iPhone, internet-enabled mobile phones existed. There was a very few websites that could be viewed on such devices and no one ever used it. The pinch-to-zoom interface of the original iPhone was revolutionary as it allowed to view almost all existing web-content on a screen size, which otherwise would have been unsuitable for the content. This feature connected an existing ecosystem to an existing device category creating a whole new experience.

Now think of automotive. The first OEMs are realizing that touch interfaces are not that great when it comes to driver distraction. Gesture control is available but limited to very specific use-cases. Speech is getting a lot better - but aside from Android or Alexa integration again limited to specific use-cases.

When I look at current Android integrations the interaction model is tailored to touchscreens. The app menu structure is more or less copied from their smartphone versions. Just like the old feature phones tried to access content from the web via an interface that was not suitable, todays application and function integration follows a usability concept that is not appropriate for the device, i.e., the vehicle.

The missing link

So far, I am missing the "pinch-to-zoom" equivalent for automotive software. If you want to make the mass market for new software functions, you need to make the user concepts mass-market suitable. Think of your parents or grand parents (especially the ones that are not engineers). What would be the entry barrier for them to use a new software function in their car? I am not only talking about the installation process but about the actual use. You definitely need to be able to adapt the user interface to the use-case of the application - as for the smartphone where you have a general purpose touchscreen. Yet, it must be such that it does not require your attention on a screen permanently (e.g. to hit a non-haptic touch button).

What the smartphone market has achieved by now, is to establish a consistent interface that structures the interaction (e.g. with gestures and notification bars etc.) while granting maximum freedom to developers so that innovation is not restricted.

The step has not been achieved in automotive yet. Everyone was jumping on the "everything must be touchscreen"-bandwagon forgetting about usability. The big touchscreen will stay in the vehicle. However, I think in the future the user-concepts will not rely on touch as main interface - at least for functions that are frequently used while driving.

You cannot require the driver to visually confirm all the time what he is pressing. You cannot require the driver to provide obvious inputs. The interface must be context-aware and intent-driven. User concepts of the smartphone world that are centered around gaining the maximum attention cannot work. You need the exact opposite - an interface targeted at the minimum required attention. There are a ton of different options - displays, sound, lighting, haptic feedback, projection, smart surfaces, I even read about thermo elements as user interface recently. I am convinced that the key to great automotive user-concepts lies in the application of voice-enabled and context-aware AI. The interesting aspect will be, how such functionality is exposed to developers via APIs in order to create a predictable and consistent user-interaction.