Impacts of visual and cognitive distractions and time pressure on pedestrian crossing behaviour: A simulator study

CBAM VGG16: An efficient driver distraction classification using CBAM embedded VGG16 architecture

Driver monitoring systems (DMS) are crucial in autonomous driving systems (ADS) when users are concerned about driver/vehicle safety. In DMS, the significant influencing factor of driver/vehicle safety is the classification of driver distractions or activities. The driver's distractions or activities convey meaningful information to the ADS, enhancing the driver/ vehicle safety in real-time vehicle driving. The classification of driver distraction or activity is challenging due to the unpredictable nature of human driving. This paper proposes a convolutional block attention module embedded in Visual Geometry Group (CBAM VGG16) deep learning architecture to improve the classification performance of driver distractions. The proposed CBAM VGG16 architecture is the hybrid network of the CBAM layer with conventional VGG16 network layers. Adding a CBAM layer into a traditional VGG16 architecture enhances the model's feature extraction capacity and improves the driver distraction classification results. To validate the significant performance of our proposed CBAM VGG16 architecture, we tested our model on the American University in Cairo (AUC) distracted driver dataset version 2 (AUCD2) for cameras 1 and 2 images. Our experiment results show that the proposed CBAM VGG16 architecture achieved 98.65% classification accuracy for camera 1 and 97.85% for camera 2 AUCD2 datasets. The CBAM VGG16 architecture also compared the driver distraction classification performance with DenseNet121, Xception, MoblieNetV2, InceptionV3, and VGG16 architectures based on the proposed model's accuracy, loss, precision, F1 score, recall, and confusion matrix. The drivers' distraction classification results indicate that the proposed CBAM VGG16 has 3.7% classification improvements for AUCD2 camera 1 images and 5% for camera 2 images compared to the conventional VGG16 deep learning classification model. We also tested our proposed architecture with different hyperparameter values and estimated the optimal values for best driver distraction classification. The significance of data augmentation techniques for the data diversity performance of the CBAM VGG16 model is also validated in terms of overfitting scenarios. The Grad-CAM visualization of our proposed CBAM VGG16 architecture is also considered in our study, and the results show that VGG16 architecture without CBAM layers is less attentive to the essential parts of the driver distraction images. Furthermore, we tested the effective classification performance of our proposed CBAM VGG16 architecture with the number of model parameters, model size, various input image resolutions, cross-validation, Bayesian search optimization and different CBAM layers. The results indicate that CBAM layers in our proposed architecture enhance the classification performance of conventional VGG16 architecture and outperform the state-of-the-art deep learning architectures.

Safety analysis of pedestrians distracted by mobile phones at street crossings: Field study in Nanjing

The growing public concern over traffic safety hazards caused by pedestrians' distracted behavior, particularly related to mobile phone usage at pedestrian crossings. Through video recording of pedestrians' street-crossing behaviors on 12 sidewalks across 9 urban road intersections in Nanjing city, 1778 valid pedestrian samples were collected. The study categorizes mobile phone use during the crossing into five distinct types: no use, voice call, screen gaze, screen gaze with operation, and listening to music with headphones. Then, the effects of gender, age and companion conditions on the mobile phone during the crossing were examined by chi-square tests. Utilizing binomial logistic and Relogit regression models, the study analyzed the impact and safety risks of distraction on crossing behavior. Additionally, a random parameters (RP) logit model with heterogeneity in means was used to investigate the determinants affecting mobile phone usage. Notably, individuals aged 18 to 30 years were identified as the random parameter, while factors such as being under 18 years old, having a companion without interaction, having a companion with interaction, weekdays, sidewalk length, and the total number of other pedestrians were identified as fixed parameters. The study also presented significant variables affecting the probability of mobile phone usage through marginal effects, highlighting the potential safety risks associated with mobile phone usage during street crossing. These findings emphasize the need for heightened pedestrian safety awareness and a reduction in distracted behaviors to enhance overall traffic safety.

Effects of smartwatch usage on Children's street-crossing behavior and visual attention allocation: An experimental study on street crossing in a real road environment

The use of smartwatches is facilitating the lives of children. However, smartwatch usage while crossing a street may lead to distractions and traffic safety hazards. The purpose of this study was to investigate the effects of smartwatch usage on the street-crossing behavior and visual strategies of children. Twenty-two children were recruited, and 220 street-crossing experiments were performed in real road environments, including at signalized and unsignalized intersections. ANOVA and Kruskal-Wallis H test calibrations were used to analyze the street-crossing behavior between the control group (who performed a task without smartwatch distraction) and the experimental group (who performed four different tasks with smartwatch distraction), and to visualize and compare the visual behavior data of children crossing the street. The results showed that children wearing smartwatches crossed the street more slowly and had a relatively narrow range of visual search, focusing on only a small area of the zebra crossing around them and the crosswalk area within their sight, whereas children who did not wear smartwatches additionally focused on areas with dangerous oncoming traffic on the left and right sides. Children wearing smartwatches were under various mental workloads and distraction levels as they performed the various distracting tasks, with listening to stories being the least distracting and least intensive, followed by answering phone calls and viewing images; a simple arithmetic operation was the most distracting and intensive task. Children while performing the distracting tasks showed a lower street-crossing speed and scanned left and right fewer times; viewing images and performing the simple arithmetic task also increased their fixation time on the smartwatches, further affecting their street-crossing behavior and thus increasing the risk of collision. At signalized intersections, children paid more attention to hazardous oncoming traffic areas and exhibited cautious crossing behaviors compared with that at unsignalized intersections. This study elucidated the characteristics of the effects of smartwatch usage on the street-crossing behavior of children and the influence pattern of the visual attention allocation and revealed the hazards of smartwatch usage on the street-crossing safety. It can contribute to improving the safety cognition of children distracted by smartwatch usage and reducing the occurrence of distracting behaviors while crossing streets. The results can provide a theoretical basis for scientific interventions and safety improvement measures.

Smartphones and rightward collisions

People tend to deviate to the right when walking through a narrow aperture (e.g., a doorway), resulting in a rightward bias in collisions. This study examines the effects of smartphone use on rightward collisions while walking. When pedestrians walk through a narrow aperture, they usually head straight to the perceived centre of the aperture, which is shifted slightly to the right, without updating the estimates. The rightward shift of the perceived centre is attributable to the rightward attentional shift in the extrapersonal space. Pedestrians using smartphones tend to fixate on the phone most of the time and thus tend not to look at their surroundings (i.e., extrapersonal space). Therefore, we predict that smartphone use will reduce rightward collisions. To test this prediction, we used a narrow-doorway task in which participants walked through a narrow doorway either with or without a smartphone. The participants with smartphones used them to perform either verbal or spatial tasks. The number of rightward collisions decreased when the participants used smartphones. The type of task had no effect on the lateral collision biases. These results were interpreted in terms of lateral attentional bias in peripersonal and extrapersonal spaces.

Deceleration parameters as implicit communication signals for pedestrians' crossing decisions and estimations of automated vehicle behaviour

Society greatly expects the widespread deployment of automated vehicles (AVs). However, the absence of a driver role results in unresolved communication issues between pedestrians and AVs. Research has shown the crucial role of implicit communication signals in this context. Nonetheless, it remains unclear how pedestrians subjectively estimate vehicle behaviour and whether they incorporate these estimations as part of their crossing decisions. For the first time, this study explores the impact of implicit communication signals on pedestrians' subjective estimations of approaching vehicle behaviour across a wide range of experimental traffic scenarios and on their crossing decisions in the same scenarios through a comprehensive analysis. Two simulator tasks, namely a natural road crossing task and a vehicle behaviour estimation task, were designed with controlled time to collision, vehicle speed, and deceleration behaviour. A novel finding is that the correlation between crossing decisions and vehicle behaviour estimations depends on the traffic scenario. Pedestrians' recognition of different deceleration behaviour aligned with their crossing decisions, supporting the notion that they actively estimate vehicle behaviour as part of their decision-making process. However, if the traffic gap was long enough, the effects of vehicle speed were the opposite between crossing decisions and estimations, suggesting that vehicle behaviour estimation may not directly impact crossing decisions when the time gap to the vehicle is large. We also found that pedestrians crossed the street earlier and estimated yielding behaviour more accurately in early-onset braking scenarios than in late-onset braking scenarios. Interestingly, vehicle speed significantly affected pedestrians' estimations, with pedestrians tending to perceive low vehicle speed as yielding behaviour regardless of whether the vehicle yielded. Finally, we demonstrated that visual cue τ̇ is a practical indicator for controlling the vehicle deceleration evidence in the experiment. In conclusion, these findings reveal in detail the role of deceleration parameters as implicit communication signals between pedestrians and AVs, with implications for road crossing safety and the development of AVs.

Rethinking infrastructure design: evaluating pedestrians and VRUs' psychophysiological and behavioral responses to different roadway designs

The integration of human-centric approaches has gained more attention recently due to more automated systems being introduced into our built environments (buildings, roads, vehicles, etc.), which requires a correct understanding of how humans perceive such systems and respond to them. This paper introduces an Immersive Virtual Environment-based method to evaluate the infrastructure design with psycho-physiological and behavioral responses from the vulnerable road users, especially for pedestrians. A case study of pedestrian mid-block crossings with three crossing infrastructure designs (painted crosswalk, crosswalk with flashing beacons, and a smartphone app for connected vehicles) are tested. Results from 51 participants indicate there are differences between the subjective and objective measurement. A higher subjective safety rating is reported for the flashing beacon design, while the psychophysiological and behavioral data indicate that the flashing beacon and smartphone app are similar in terms of crossing behaviors, eye tracking measurements, and heart rate. In addition, the smartphone app scenario appears to have a lower stress level as indicated by eye tracking data, although many participants do not have prior experience with it. Suggestions are made for the implementation of new technologies, which can increase public acceptance of new technologies and pedestrian safety in the future.