novices vs. expert users

who should test our HMI?

should we use people with no prior knowledge of the system or test people with system knowledge?

if you want to capture first impressions you have to test "naive" participants. if you want to test system efficiency and effectiveness you should test participants with some system knowledge.

I am not sure if you can do both. for example, in one on-road study (see video), the owner of the car had no knowledge of a navigation feature and that resulted in 3 times more eyes-of-the-road time in comparison with his 2nd attempt that he knew about the feature. significant difference on performance.

because I really love on-road studies, and because I don't want people learning the system while they are driving, I use people with system knowledge or I train them before the task. I try to capture some, limited, first impressions during system setup (e.g. phone pairing and home/work address setup).

what is your approach ? have you got the resources to run longitudinal studies?

Data collection and analysis as part of a personal study conducted in the summer of 2018. Many thanks to Scott Hodgins and Jon Ward at Tobii Pro

HUD and visual attention

If we want eyes on the road and hands on the steering wheel, we must provide head-up-display (HUD) to drivers.

HUD is not ideal for all the use cases, but it can be extremely powerful on certain tasks. For example, with HUD I had my eyes on the road and hands on steering wheel when I had to change my radio station from BBC R2 FM to DAB Jazz Radio.

Doing the same task with the iDrive was very efficient (14 seconds), but I had my eyes 5 seconds away from the road, in total, during the task. Doing the same task with the touchscreen was way too difficult both on task time and eyes off the road (I will post some data in the future).

HUD is usually an expensive optional for the customer to purchase. My opinion is that automotive manufacturers should invest more into HUD technologies (either windscreen or combiner) and start offering HUD as standard equipment.

*data collection and analysis for a personal study

BMW voice interaction

everyone is asking for "intuitive" HMI, for systems that will "surprise and delight" the user.

but, what do these words really mean to the people that design the system? how can you convert these abstract terms into requirements?

I am not an abstract thinker, I like tangible outcomes, prototypes and examples. So here is a good example:

The BMW voice system it was a "surprise and delight" when I asked a question about the fuel level.

I asked a question, very vague and ill-defined, and the system provided: a) current range, b) distance to my destination as it was entered to the navigation system and c) it informed/assured me that it was within range.

This is what I call "surprise and delight"