Measuring Driver Perception: Combining Eye-Tracking and Automated Road Scene Perception

The Effects of Auditory Working Memory Task on Situation Awareness in Complex Dynamic Environments: An Eye-movement Study

OBJECTIVE

This study investigated the effect of auditory working memory task on situation awareness (SA) and eye-movement patterns in complex dynamic environments.

BACKGROUND

Many human errors in aviation are caused by a lack of SA, and distraction from auditory secondary tasks is a serious threat to SA. However, it remains unclear how auditory working memory tasks affect SA and eye-movement patterns.

METHOD

Participants (n = 28) were randomly allocated to two groups and received different periods of visual search training (short versus long). They subsequently completed a situation awareness measurement task in three auditory secondary task conditions (without secondary task, auditory calculation task, and auditory 2-back task). Eye-movement data were collected during the situation awareness measurement task.

RESULTS

The auditory 2-back task significantly reduced overall SA, Level 1 SA, dwell times, and total percentage of fixation time on task-related areas of interests in the SA measurement task. Overall SA and Level 3 SA were not reduced by the auditory 2-back task in individuals in the longer visual search training time condition.

CONCLUSION

Auditory working memory load impairs SA in the perception and projection stage; however, greater experience can overcome impairment of SA in the projection stage.

APPLICATION

This study provided possible approaches to preventing loss of SA: (1) improving crew members' communication skills to ensure the accurate and clear transmission of information, reducing the difficulty of processing information, and (2) providing targeted cognitive training tailored to each pilot's level of experience.

Analyzing Factors Influencing Situation Awareness in Autonomous Vehicles-A Survey

Autonomous driving of higher automation levels asks for optimal execution of critical maneuvers in all environments. A crucial prerequisite for such optimal decision-making instances is accurate situation awareness of automated and connected vehicles. For this, vehicles rely on the sensory data captured from onboard sensors and information collected through V2X communication. The classical onboard sensors exhibit different capabilities and hence a heterogeneous set of sensors is required to create better situation awareness. Fusion of the sensory data from such a set of heterogeneous sensors poses critical challenges when it comes to creating an accurate environment context for effective decision-making in AVs. Hence this exclusive survey analyses the influence of mandatory factors like data pre-processing preferably data fusion along with situation awareness toward effective decision-making in the AVs. A wide range of recent and related articles are analyzed from various perceptive, to pick the major hiccups, which can be further addressed to focus on the goals of higher automation levels. A section of the solution sketch is provided that directs the readers to the potential research directions for achieving accurate contextual awareness. To the best of our knowledge, this survey is uniquely positioned for its scope, taxonomy, and future directions.

The Relationship between Stress Levels Measured by a Questionnaire and the Data Obtained by Smart Glasses and Finger Pulse Oximeters among Polish Dental Students

Stress is a physical, mental, or emotional response to a change and is a significant problem in modern society. In addition to questionnaires, levels of stress may be assessed by monitoring physiological signals, such as via photoplethysmogram (PPG), electroencephalogram (EEG), electrocardiogram (ECG), electrodermal activity (EDA), facial expressions, and head and body movements. In our study, we attempted to find the relationship between the perceived stress level and physiological signals, such as heart rate (HR), head movements, and electrooculographic (EOG) signals. The perceived stress level was acquired by self-assessment questionnaires in which the participants marked their stress level before, during, and after performing a task. The heart rate was acquired with a finger pulse oximeter and the head movements (linear acceleration and angular velocity) and electrooculographic signals were recorded with JINS MEME ES_R smart glasses (JINS Holdings, Inc., Tokyo, Japan). We observed significant differences between the perceived stress level, heart rate, the power of linear acceleration, angular velocity, and EOG signals before performing the task and during the task. However, except for HR, these signals were poorly correlated with the perceived stress level acquired during the task.

Recognition of Drivers’ Activity Based on 1D Convolutional Neural Network

Background and objective: Driving a car is a complex activity which involves movements of the whole body. Many studies on drivers’ behavior are conducted to improve road traffic safety. Such studies involve the registration and processing of multiple signals, such as electroencephalography (EEG), electrooculography (EOG) and the images of the driver’s face. In our research, we attempt to develop a classifier of scenarios related to learning to drive based on the data obtained in real road traffic conditions via smart glasses. In our approach, we try to minimize the number of signals which can be used to recognize the activities performed while driving a car. Material and methods: We attempt to evaluate the drivers’ activities using both electrooculography (EOG) and a deep learning approach. To acquire data we used JINS MEME smart glasses furnished with 3-point EOG electrodes, 3-axial accelerometer and 3-axial gyroscope. Sensor data were acquired on 20 drivers (ten experienced and ten learner drivers) on the same 28.7 km route under real road conditions in southern Poland. The drivers performed several tasks while wearing the smart glasses and the tasks were linked to the signal during the drive. For the recognition of four activities (parking, driving through a roundabout, city traffic and driving through an intersection), we used one-dimensional convolutional neural network (1D CNN). Results: The maximum accuracy was 95.6% on validation set and 99.8% on training set. The results prove that the model based on 1D CNN can classify the actions performed by drivers accurately. Conclusions: We have proved the feasibility of recognizing drivers’ activity based solely on EOG data, regardless of the driving experience and style. Our findings may be useful in the objective assessment of driving skills and thus, improving driving safety.