Eyes on the road

x Sponsored content on Research Live and in Impact magazine is editorially independent.
Find out more about advertising and sponsorship.

As car manufacturers are incorporating more technologies into their vehicles, and connected cars are changing the in-vehicle driver experience, it has become vital to test how these technologies affect drivers and balance the conflicting interests of people wanting to be connected but also to be safe.

So market research agency Eye Square has been working with car brands and developing means of tracking drivers’ reactions with the help of implicit research techniques. The particular challenge has been in using multiple methods in combination with each other, to get more in-depth understanding of how connected technology impacts on driver safety and awareness.

The implicit methods used include eye tracking; facial emotion analysis; and physiological measures through EEG (electroencephalography) where brain activity shows a driver’s emotional state and EDA (electrodermal activity) which shows arousal.  

There are considerable physical constraints that the in-car environment presents researchers and any wearable devices must not impede the driver in any way.

“Driving is our primary task and then we have secondary and territory tasks. The primary task must have the highest security measures so they don’t interfere with anything and the driver should not be distracted by them or not know what they do. Eye tracking is the most obtrusive one as you wear them on your face,” says Philipp Leppert, partner and COO, Eye Square.

“It is important there are no cables. The eye tracker we use now is so light and has no restrictions in the peripheral vision. We see a lot on the left and right where we are very sensitive to movement so even though drivers are looking straight they almost have 180 degree vision,” he adds.

Through eye tracking the average share of dwell time on any given area within the car, as well as through the windscreen can be determined.

The implicit measures include facial expressions looking at variations on two major muscles – one between the eyes and one for mouth – to differentiate between smiling or frowning – “it’s an automatic detector of your valence – positive or negative”, says Leppert. However care is needed in the interpretation as negative could mean frustration but also concentration, or maybe the driver was just blinded by the sun.

Similarly skin conductance is a very sensitive measure of arousal – “it could be stress or positive because you’re driving fast and it’s fun”, says Leppert. “It gives us a second by second indication of stimuli but also a long term curve to see how your body reacts to the whole situation.”

The overall connected-car driving experience can be measured by combining these implicit methods with the physical ones such as which button they pressed (which the researchers can determine from the car and also from observation.

But one of the most challenging aspects is ensuring all the readings are synchronised so that the data from different sources can be collated correctly.

“It’s quite demanding for our analysis team. Firstly they record separately but it’s all synchronised with the same atomic clock and then we aggregate them; the camera can tell you where the car was and what the situation was. We build sequences in driving – these maybe tasks such as ‘please drive to destination X’ and then we make sure the people are communicating with car, such as having to programme the navigation system. These are situations where we objectively rate the situation on how cognitively demanding the primary task is. So if it’s demanding – driving over 100 m/hr for instance – the attention shouldn’t be diverted from the road, whereas if they are waiting at a red light it’s okay to look at the radio,” explains Leppert.

And to ensure there’s a control, all the neuro and implicit measures are also recorded for the driver in a non-driving part – these are the baseline recordings Eye Square does in every car with every person because each individual has different physiological levels.