Introduction

The occlusion technique: a procedure to assess the HMI of in-vehicle information and communication systems

In-vehicle information systems (IVIS) demand driver attention and lead to driver distraction and inattention, impairing driving safety. The Box Task (BT) quantifies these secondary task demands concerning visual and manual aspects of the driving task. Combined with the tactile Detection Response Task (tDRT), it also measures cognitive demands. The standard method for assessing secondary task demand is the occlusion method. The present study compares the combined BT+tDRT method concerning various secondary task demands with the standardized occlusion method. In a laboratory study, participants completed artificial and realistic IVIS tasks with BT+tDRT or occlusion. Results revealed that BT+tDRT was able to distinguish different demands of visual, manual, and cognitive tasks, while the occlusion method was only able to distinguish different visual-manual IVIS tasks. Based on the results, we conclude that the combination of BT+tDRT captures the effects of IVIS on driving safety more precisely and comprehensively than does the occlusion method. Future research should compare the findings from this study with data from field studies. Also, the use of BT+tDRT should be further standardized.

Instantaneous consumption displays (ICDs) can be used as central information source to perceive the energy efficiency of manoeuvre-level driving. A key question is whether drivers who use ICDs can accurately derive energy efficiency differences of different driving strategies based on ICDs. There is reason to assume that drivers' consumption judgements may be biased, similar to driving-related phenomena like the time-saving bias. Therefore, the aim of the present research was to examine drivers’ accuracy in deriving average consumption from dynamic ICD sequences. Participants viewed videos of a schematic ICD in a controlled experiment where the maximum instantaneous consumption systematically varied over time. Participants (N = 55) overestimated the average consumption values. The empirical ranking of the sequences did significantly correlate with the heuristic but not with the correct efficiency ranking. The current study incorporated multilevel modelling due to the nested structure of the data. The estimation difference was greater with higher peak height and shorter peak duration. The effect of peak height on estimation difference weakened with longer peak duration. In sum, the results indicate that ICDs can create biased perceptions of energy efficiency and that drivers seem to use simplifying heuristics. Knowledge and affinity for technology interaction appear to relate to biased estimations, whereas the intensity of prior experience with consumption displays seems irrelevant. Further studies should test other interfaces with debiasing potential such as manoeuvre-based aggregation or fading-trace approaches. Moreover, studies are needed that enable modelling of the effects of more natural temporal-spatial visual attention distribution (e.g. in a driving simulator setting).

Several tools have been developed over the past twenty years to assess the degree of driver distraction caused by secondary task engagement. A relatively new and promising method in this area is the box task combined with a detection response task (BT + DRT). However, no evaluation regarding the BT's sensitivity currently exists. Thus, the aim of the present study was to evaluate the BT + DRT by comparing its sensitivity to the sensitivity of already established methods. Twenty-nine participants engaged in several artificial and realistic secondary tasks while either performing the BT + DRT, the Lane Change Test (LCT), or driving through a simple course in a simulator. The results showed that the BT parameters (especially the standard deviation of box position and size) were sensitive to differences in demand across the visual-manual secondary tasks. This was comparable to what was found with the LCT. Surprisingly, the BT performance measures were more sensitive than those of the driving simulation task. The BT + DRT also captured cognitive distraction effects with the integration of the DRT. Hence, the BT + DRT could be a cost-effective method to assess in-vehicle system demand. However, further investigations are necessary to better understand the potential of the BT method.

The ability to assess the current attentional state of the driver is important for many aspects of driving, not least in the field of partial automation for transfer of control between vehicle and driver. Knowledge about the driver’s attentional state is also necessary for the assessment of the effects of additional tasks on attention. The objective of this paper is to evaluate different methods that can be used to assess attention, first theoretically, and then empirically in a controlled field study and in the laboratory.

Six driving instructors participated in all experimental conditions of the study, delivering within-subjects data for all tested methods. Additional participants were recruited for some of the conditions. The test route consisted of 14 km of motorway with low to moderate traffic, which was driven three times per participant per condition. The on-road conditions were: baseline, driving with eye tracking and self-paced visual occlusion, and driving while thinking aloud. The laboratory conditions were: Describing how attention should be distributed on a motorway, and thinking aloud while watching a video from the baseline drive.

The results show that visual occlusion, especially in combination with eye tracking, was appropriate for assessing spare capacity. The think aloud protocol was appropriate to gain insight about the driver’s actual mental representation of the situation at hand. Expert judgement in the laboratory was not reliable for the assessment of drivers’ attentional distribution in traffic. Across all assessment techniques, it is evident that meaningful assessment of attention in a dynamic traffic situation can only be achieved when the infrastructure layout, surrounding road users, and intended manoeuvres are taken into account. This requires advanced instrumentation of the vehicle, and subsequent data reduction, analysis and interpretation are demanding. In conclusion, driver attention assessment in real traffic is a complex task, but a combination of visual occlusion, eye tracking and thinking aloud is a promising combination of methods to come further on the way.

The MARS (Masking Action Relevant Stimuli) method assesses information demand for dynamic stimuli while driving. An action relevant stimulus is masked and the driver presses a button to unmask the stimulus for a limited period. We interpreted button presses as information demand. Following our previous research (Rittger, Kiesel, Schmidt, & Maag, 2014), the current study further evaluates the method. We applied the MARS method to a dynamic in-vehicle display containing recommendations from a traffic light assistant. In a driving simulator, drivers approached intersections with different traffic light phasing. The display either presented simple or complex information. In half of the drives, the participants used the MARS method. The study had a full within subjects design and fixations were recorded in all conditions. The results showed that the information demand varied according to the information in the display and the traffic light phase. A comparison of button presses with fixations showed that one unmasking interval came along with one fixation on the display. As a conclusion, the MARS method can distinguish between conditions with high and low information demand for the display. Button presses relate to fixations on the display. Hence, the MARS method is a promising tool to assess the information demand in dynamic environments and can be applied as an extension or alternative for eye tracking.

We present the novel MARS (Masking Action Relevant Stimuli) method for measuring drivers’ information demand for an action relevant stimulus in the driving scene. In a driving simulator setting, the traffic light as dynamic action relevant stimulus was masked. Drivers pressed a button to unmask the traffic light for a fixed period of time as often as they wanted. We compared the number of button presses with the number of fixations on the traffic light in a separate block using eye tracking. For the driving task, we varied the road environment by presenting different traffic light states, by adding a lead vehicle or no lead vehicle and by manipulating the visibility of the driving environment by fog or no fog. Results showed that these experimental variations affected the number of button presses as dependent measure of the MARS method. Although the number of fixations was affected qualitatively similar, changes were more pronounced in the number of fixations compared to the number of button presses. We argue that the number of button presses is an indicator for action relevance of the stimulus, complementing or even substituting the recording and analyses of gaze behaviour for specific research questions. In addition, using the MARS method did not change dynamic driving behaviour and driving with the MARS method was neither disturbing, nor difficult to learn. Future research is required to show the generalisability of the method to other stimuli in the driving scene.