The investigation of interaction mechanisms with in-vehicle infotainment systems (IVIS)

Driver distraction has been a concern for the automotive industry and for researchers in the area of driving Human Factors for decades. The increasing use of touch screens in particular as a means of enabling drivers access to infotainment functions are providing more opportunities for visual demand, leading to a higher number of off-road glances and a detrimental impact on driver performance; potentially causing a risk to the driver themselves as well as other road users. This thesis reports on five separate driving simulator studies which focus on identifying various design characteristics of in-vehicle touch screens that contribute to off-road glances. It also aims to investigate the interaction mechanisms used by drivers whilst driving that achieve non-visual interaction with the in-vehicle display. In study one, twenty participants took part in a driving simulator study which investigated the effects of various design characteristics of an in-vehicle touch screen (button location, size and contrast level) on driver glance behaviour whilst driving. The results of this study showed that larger buttons resulted in fewer off-road glances. The results also highlighted that across all conditions the proportion of zero glance interactions was between 10 and 45 per cent showing that participants were able to perform the button selection task without taking their eyes off the road at times. In study two, the effects of the number of items displayed on an in-vehicle touch screen on driver glance behaviour whilst driving were investigated in a simulator study with sixteen participants. The results showed that a higher number of items presented on the touch screen resulted in more glances being made away from the road. The proportion of zero glance interactions was between 12 and 24 per cent, varying based on the number of buttons displayed which indicated that participants took their eyes off the road to look at the touchscreen for most interactions, but vision was not needed for all successful target selections. Observations made of the behaviours of the participants when interacting with the touch screen and feedback they gave helped in identifying two possible interaction mechanisms for non-visual interaction; namely, peripheral vision and muscle memory. In study three, twenty-five participants completed a driving simulator study which aimed to identify the interaction mechanisms adopted by those who were able to perform the button selection task non-visually. Two interaction mechanisms (peripheral vision and muscle memory) were tested with different sized buttons presented on the touch screen display. The results showed that participants were able to interact with larger buttons using only peripheral vision which resulted in them being able to keep their eyes on the road ahead. In contrast, muscle memory appeared to have little impact on results. In study four, the relationships between driving complexity and the two interaction mechanisms investigated in study three were examined in another driving simulator study that involved twenty participants. The two interaction mechanisms were also compared to foveal vision to investigate how successful participants were in completing a button selection task on an in-vehicle touchscreen. The results of the study showed that although driving complexity did not impact participants’ performance, there was a considerable difference between the interaction mechanisms. In particular, a muscle memory-based interaction resulted in considerably worse task and driving performance compared to peripheral and foveal vision interactions. In the final study, twenty participants took part in a driving simulator study in which a peripheral vision friendly interface (based on the design elements identified from the literature and the first two studies) was tested with either peripheral vision interaction or natural interaction. In the natural condition, participants were asked to interact with the touch screen as they would in a real on-road driving environment. The results of this study showed that the driving performance when interacting with the display using peripheral vision showed similarities to the performance when participants were interacting with the touch screen naturally. In conclusion, in contrast to previous literature on the investigation of the role of peripheral vision in a driving context, this thesis investigates the role of peripheral vision in interaction with in-vehicle displays. The work highlights that peripheral vision as an interaction mechanism with in-vehicle touch screens can result in successful ‘blind interactions’ allowing drivers to keep their eyes on the road whilst driving to decrease the possibility of driver distraction caused by visual demands placed on the driver. It also shows that there are clear design elements of in-vehicle touch screens that can be used in aiding peripheral vision interaction, for example the number of items and the location of the items displayed on the screen as well as the contrast levels of the display. Future work should focus on testing in-vehicle infotainment systems present in vehicles to identify which aspects could be redesigned to support peripheral vision interaction in a real-driving environment