A Nonintrusive and Real-Time Classification Method for Driver’s Gaze Region Using an RGB Camera

Driver distraction has become a leading cause of traffic crashes. Visual distraction has the most direct impact on driving safety among various driver distractions. If the driver’s line of sight deviates from the road in front, there will be a high probability of visual distraction. A nonintrusive and real-time classification method for driver’s gaze region is proposed. A Multi-Task Convolutional Neural Network (MTCNN) face detector is used to collect the driver’s face image, and the driver’s gaze direction can be detected with a full-face appearance-based gaze estimation method. The driver’s gaze region is classified by the model trained through the machine learning algorithms such as Support Vector Machines (SVM), Random Forest (RF), and K-Nearest Neighbors (KNN). The simulated experiment and the real vehicle experiment were conducted to test the method. The results show that it has good performance on gaze region classification and strong robustness to complex environments. The models in this paper are all lightweight networks, which can meet the accuracy and speed requirements for the tasks. The method can be a good help for further exploring the visual distraction state level and exert an influence on the research of driving behavior.

Effects of automation trust in drivers’ visual distraction during automation

With ongoing improvements in vehicle automation, research on automation trust has attracted considerable attention. In order to explore effects of automation trust on drivers’ visual distraction, we designed a three-factor 2 (trust type: high trust group, low trust group) × 2 (video entertainment: variety-show videos, news videos) × 3 (measurement stage: 1–3) experiment. 48 drivers were recruited in Dalian, China for the experiment. With a driving simulator, we used detection-response tasks (DRT) to measure each driver’s performance. Their eye movements were recorded, and automation-trust scale was used to divide participants into high trust group and low trust group. The results show that: (1) drivers in the high trust group has lower mental workload and paid more attention to visual non-driving-related tasks; (2) video entertainment also has an impact on distraction behavior, variety-show videos catch more attention than news videos. The findings of the present study indicate that drivers with high automation trust are more likely to be involved in non-driving-related visual tasks.

Distraction in the Driving Simulator: An Event-Related Potential (ERP) Study with Young, Middle-Aged, and Older Drivers

Distraction is a major causal factor of road crashes, and very young and older drivers seem to be particularly susceptible to distracting stimuli; however, the possibilities of exploring the causes for increased distractibility of these groups in real traffic seem to be limited. Experiments in a driving simulator are a good choice to eliminate the risk for crashes and to present highly standardized stimulus combinations. In the present study, 72 subjects from four age groups completed a driving task that required occasional responses to the brake lights of a car in front. In addition, in certain experimental conditions, subjects had to respond to distracting visual or auditory stimuli. In addition to behavioral data, electrophysiological correlates of stimulus processing were derived from the electroencephalogram (EEG). In the two older groups, braking response times increased even in a simple task condition when visual distraction stimuli occurred. In more complex task conditions braking response times increased with acoustic and visual distractors in the middle-aged group as well. In these complex task conditions braking error rates, especially the missing of braking reaction in favor of the distracting task, increased under visual distraction with increasing age. Associated with this, a reduced P3b component in the event-related potential indicated an unfavorable allocation of mental resources. The study demonstrates the potential of driving simulators for studying effects of distraction, but also their limitations with respect to the interpretability of the results.

A combined Nutri-Score and ‘Eco-Score’ approach for more nutritious and more environmentally friendly food choices? Evidence from a Belgian consumer experiment.

The application of Nutri-Score on food products is ubiquitous throughout Europe and studies demonstrating its potential to stimulate healthier food choices are accumulating. At the same time, there remains a strong need to evenly harmonize and activate the communication of environmental impacts on food products, in synergy with the Nutri-Score. This brings up the question of whether the potential of Nutri-Score could be expanded to a similar ‘Eco-Score’ and equivalently encourage more environmentally friendly food choices. The present study investigated the effect of a combined NutriScore and Eco-Score on the nutritional quality and environmental impact of consumers’ food choices. This effect was compared to, on the one hand, dietary recommendations (both general and specific) and on the other hand, more detailed impact tables. Since visual distraction often plays a role in informative persuasion, the treatments were evaluated subject to different levels of distraction. A randomized control trial was conducted with a representative sample of 805 Belgian consumers in a mock-up E-grocery environment. Respondents were randomly allocated to treatments in which they were asked to hypothetically buy ingredients to prepare one meal. An average nutritional quality index (NQI) and environmental impact index (EII) of the selected baskets were calculated to evaluate outcomes. We find that a joint Nutri-Score and Eco-Score label improves the NQI but not the EII. The general- and specific recommendation as well as the detailed information also improved the NQI. However, the specific recommendation was the only treatment that also improved the EII. We find mild indications that the effectiveness of Nutri-Score is affected by the nearby presence of food product images. This study provided some first evidence and support for the use of dual Nutri-Score – Eco-Score label to induce transitions towards healthier and more sustainable diets. We also find that recommendations outside the classic Front-Of-Package label framework are also a promising way to realize such transition. However, the effect in real E-groceries and on longer terms remains to be explored.