Sage Journals: Discover world-class research

Abstract

Anxiety is a common emotion of driver, which always affects the safety of driving. Eye movement characteristics can be used to understand the true emotion state of human beings. It is of great significance to study the law of eye movement for realizing active vehicle safety warning and human–machine cooperation. In this article, anxiety-induction experiment, real-vehicle driving experiments, and virtual driving experiments were designed and used to obtain the eye movement data of female novice extroversion driver under calm and anxiety, and mathematical statistics analysis was made on the fixation count, fixation duration, and visit duration in the area of interest within the driver horizon. The results showed that there are significant differences in fixation count and fixation duration of drivers ( $p < 0.05$ , p is the accompanying probability), and the main effect of emotion is significant $(p < 0.05)$ .Compared with the situation of calm, fixation area, fixation count, and fixation duration of drivers under anxiety were more focused on the middle area, the fixation count and visit duration on the left area were relatively more, the fixation duration on the right area was relatively longer, and anxiety was more likely to cause driver’s attention bias.

Keywords

Traffic safety driver anxiety eye movement fixation

Introduction

With the rapid development of global automobile industry and the popularization of private cars, casualties and property losses caused by traffic congestion¹ and traffic accidents have been increasing year after year. The analysis results from relevant experts on traffic accidents show that more than 90% of the occurred traffic accidents are caused by human,² in which more than 70% are caused by driver. During driving, the proportion of driver using vision to perceive various traffic information is as high as 80%.³ The eye movement law in the driver’s visual perception process must be studied in order to prevent the occurrence of traffic accidents.

Due to the great progress in the field of artificial intelligence and computer technology, the visual search mode of driver was studied by traffic scholars of relevant field in different road environments such as sharp turn section of mountain road,⁴ road construction section,⁵ tunnel entrance section,⁶ tunnel lighting environment,⁷ traffic sign and billboard,^8–10 roundabout,¹¹ curved road,^12,13 grassland road,¹⁴ and urban road.¹⁵ In terms of the relationship between emotion and visual search pattern, Dickins et al.^16–18 found that eye gaze time was lengthened when learning under the negative emotion. The effect of anxiety on attention bias was studied using eye tracking technology by Byrow and Holas,^19,20 and the results showed that eye movement was more focused on sad and happy faces than neutral faces. Individuals with high trait anxiety searched angry faces faster significantly than smiling faces. The effects of anxiety on attention and gaze behavior in the field of aviation were studied by Allsop and Gray,²¹ which found that anxiety increased the randomness of external stay time and visual scanning. Pasupa et al.²² found that a more accurate prediction was provided using binocular motion information and image features together to determine user emotion than using image features alone based on the improved prediction method. During driving, the relationship between emotion and attention was studied through risk perception by Zimasa and El Haj,^23,24 and it was found that negative emotion was harmful to driving safety than positive or neutral emotion, but emotion was not be materialized. Schmidt-Daffy and Yao^25,26 found that lowering the vehicle speed can reduce the driver’s feeling of impulses associated with danger in many dynamic models of driving behavior and emotion symptoms were attributed to fear and anxiety. Taylor et al.²⁷ found through questionnaire survey that drivers with high anxiety were more likely to make mistakes. From previous studies, there was a common finding that anxiety can narrow attention, which led to neurasthenia under stringent state, important information was ignored at the periphery and driving risk increased.²⁸ In total, 168 drivers were tested by Pourabdian and Azmoon²⁹ by utilizing the Manchester Driving Behavior Questionnaire and the Spielberg State-Trait Anxiety Scale to analyze the relationship between trait anxiety level and incorrect driving behavior, and it was found that trait anxiety levels were prone to lead to distraction and forgetfulness for driver. Wester et al.³⁰ found that driver’s fixation frequency increased in the central area and the fixation frequency on both sides decreased under high anxiety situation and attention was narrowed. However, only the index of horizontal fixation was chosen and other eye movement indexes were not introduced in this study. In addition, using a racing game in the experiment was a certain difference from the real scene.

The previous study on driver’s eye movement characteristics mainly focused on different road environments and rarely combined emotion with road environment from the relevant scholars at home and abroad. The emotion was not embodied, which affected the accuracy of result to some extent. In view of this, anxiety-induction experiment, real-vehicle driving experiment, and virtual driving experiment were designed and used to pre-collect the eye movement indexes of female novice extroversion driver. The eye movement law of driver under calm and anxiety was obtained through the analysis of the collected data, which provided the theoretical basis for the realization of active vehicle safety driving warning system and ensures the safety of driving.

Methods

Index selection of eye movement

During driving, there are three basic eye movement patterns, namely, fixation, saccade, and blink. Common characteristic indexes are shown in Table 1.

Table 1.

Common characteristic indexes of three eye movement patterns under calm and anxiety.

Eye movement pattern	Characterization index
Fixation	Fixation count, fixation duration, fixation sequence, sum of fixation duration, visit duration
Saccade	Saccade duration, saccade range, average speed of saccade, average acceleration of saccade
Blink	Blink duration, blink interval, blink frequency

Referring to research results from predecessors in the visual direction of driver,³¹ based on the specific function of Tobii head-mounted eye tracker, the fixation count, fixation duration, and visit duration were selected (shown as in Table 2 for details); driver visual characteristics under calm and anxiety was comparatively studied.

Table 2.

Common eye movement indexes under calm and anxiety.

Eye movement indexes	Description
Fixation count	The spatial focus of driver’s attention over a period of time is related to the number but has nothing to do with the depth of information that the observer needs to process. The frequency of fixation is also high in the area with high information ratio, which reflects AOI for driver.
Fixation duration	The duration that the center position of visual axis remains constant during fixation, represents the time for processing the hazard-related information and reflects the difficulty of extracting information.
Visit duration	The statistics of duration for accessing to an AOI, that is, the sum of each fixation duration in the area and the saccade time between every two points, is the comprehensive expression of the information amount in the area and the difficulty degree of information processing.

AOI: area of interest.

Participants

Related research on driver characteristics showed that women were more prone to anxiety than men.^32,33 Novice drivers had lower performance in risk-perception tasks and were more likely to be involved in road traffic accidents.^34–36 In view of this, a total of 40 female novice extroversion drivers were selected for eye movement data acquisition under calm and anxiety. The driving propensity of subjects is determined by Driving Propensity Questionnaire,³⁷ and the specific performance is shown in Table 3. The biological age ranges from 21 to 45 years and driving age ranges from 1 to 23 years. Driving experience is divided with the standard on driving mileage of 10,000 km.³⁸ All subjects have normal hearing, vision or corrected vision, and normal color vision.

Table 3.

Driving propensity type and its performance.

Driving propensity type	Performance
Introversion	Steady, prudent, stable attention and difficult to shift, not easy to generate risk-taking motivation, easy to drive at low speed, fear of traffic accidents.
Middle type	Obey the traffic rules strictly, observe thoughtfully, consider the complex traffic situation, more calm, self-control.
Extroversion	Sensitive, active, prone to generate risk-taking motive, rapid response, impetuous in the case of complex traffic, not a careful observation.

Experiment conditions

Emotion-induction materials

Anxiety-induction materials are the fundamental components of ensuring anxiety to be stimulated and are the basis to ensuring validity of subsequent eye movement parameters. Anxiety-induction materials used in this experiment mainly include International Affective Picture System (IAPS) and Chinese Affective Picture System (CAPS). IAPS are the authoritative materials of emotion induction internationally. CAPS are the emotion-induction materials that adapt to Chinese unique social and cultural backgrounds. According to the different sensory channels of materials, emotion-induction materials used in this experiment include visual materials (words and pictures with anxiety, changes in driving environment brightness, etc.), auditory materials (noisy and irregular audio, etc.), olfactory materials (cigarettes, durian, licorice, etc.), gustation materials (dark chocolate, etc.), and multi-channel stimulus materials (video with anxiety, etc.). A part of the emotion materials is shown in Figure 1.

Figure 1.

Partial materials of anxiety induction: (a) partial visual stimulus of anxiety and (b) anxiety face pictures.

Instruments

Experiment instruments include comprehensive experiment vehicle for road traffic (equipped with 32-line laser radar, laser ranging sensor, SG299GPS non-contact multi-function speedometer, CTM-8A non-contact multi-function speedometer, vehicle recorder, Tobii eye tracker, WTC-1 pedal power manipulator, high-definition camera, laptop, etc.), high simulation platform of virtual vehicle driving, camera Tobii Studio software for data analysis, recorder, IR-Marker, wedge foam, and so on. A part of the experiment instruments is shown in Figure 2.

Figure 2.

Experiment instruments.

Driving route

The road sections of Beijing Road–Renmin West Road–Nanjing Road–Xincun West Road in Zibo city (shown as in Figure 3) were selected as the driving route under good weather and road conditions. The whole length of the line is about 5.6 km. The subjects of this article are the novice drivers. The stability of driver’s calm state will be affected by the large or complex traffic flow. Therefore, the driving experiment under driver’s calm state was carried out during the off-peak time on Saturday and Sunday morning, driving experiment under drivers’ anxiety was was carried out at the early peak and late peak time from Monday to Friday. A virtual driving road of one-way dual lane was set according to the research requirement, and different virtual driving scenes were set in the experiment section (shown as in Figure 4), including unimpeded road, road maintenance, traffic accident, curve, and queuing waiting of intersection, and so on. There were both common environments (no artificial barriers) and special circumstances (pre-set obstacles at specific locations in the driving environment), which can stimulate or enhance anxiety, to ensure research is of universal significance.

Figure 3.

Driving route applied in real vehicle.

Figure 4.

Part of virtual driving scenes.

Procedure

The experiments are divided into real-vehicle driving and virtual driving. The data obtained from real-vehicle experiment are more similar to natural driving data, but the experiment organization is more difficult and the cycle is longer. So, it is difficult to obtain large amounts of eye movement data. Relatively, virtual driving experiment is simple, the cycle is short, the experiment organization and data collection process are easy, and experiment data can be supplied effectively.

Preparation

Before experiment, the organizers fixed the IR-Marker of Tobii eye tracker on the front windshield according to the 4 × 6 matrix first. The rearview mirror, steering wheel, dashboard, and so on were also placed in accordance with the IR-Marker placement requirements. The distribution diagram of IR-Marker is shown in Figure 5. Since the driver sat on the left side of vehicle while driving, the IR-Marker was integral to the left. The wedge foam was used to offset the tilt angle of the front windshield in order to ensure that each IR-Marker can face the driver directly, so the driver’s eyeball could be better captured with the eye tracker in the experiment. The eye tracker was corrected to ensure that higher quality eye movement data can be obtained during the experiment.

Figure 5.

Distribution diagram of IR-Markers.

Emotion-induction and driving experiments

Anxiety is divided into two forms: state anxiety and trait anxiety. Anxiety in this study includes state anxiety, generated by driver under the stimulation of experiment conditions during driving, and their own unique trait anxiety. The driving experiment process under calm and anxiety is shown in Figure 6.

Figure 6.

Driving experiments design.

Evaluation of anxiety-induction effect

The anxiety-induction effect affects the accuracy and reliability of the results directly. During driving, the behaviors of subjects such as facial expressions, driving condition, driving speed, and pedaling force were recorded using instruments such as camera, speedometer, and pedal power manipulator. After driving, the subjects watched the video playback during their driving process and reported the experience of emotion intensity at different moments to the experiment organizers. Fragments of eye tracker shooting were extracted, respectively, under calm and anxiety. The Beck Anxiety Scale, the Self-Rating Anxiety Scale, and the subjects’ physiological characteristics such as facial expression, speech signal, and behaviors during driving were comprehensive used to evaluate the anxiety-induction effect. The anxiety selected in this study was moderate and severe anxiety of drivers who scored more than 25 (Beck Anxiety Scale) and greater than 59 (Self-Rating Anxiety Scale) under the stimulation of experiment conditions (shown as in Table 4), respectively.

Table 4.

Basis of anxiety degree on Baker Anxiety Scale and Self-Rating Anxiety Scale.

Scale type	Anxiety degree
Scale type	No anxiety	Mild anxiety	Moderate anxiety	Severe anxiety
Beck Anxiety Scale	$< 15$	$15 - 25$	$26 - 35$	$> 35$
Self-Rating Anxiety Scale	$< 50$	$50 - 59$	$60 - 69$	$> 69$

In accordance with the above experiment steps, 40 female novice extroversion drivers were organized to carry out real-vehicle and virtual driving experiments, and the data were collected by eye tracker. After calm and anxiety were stimulated, respectively, subjects drove on specific road section at each experiment. In the real-vehicle and virtual driving experiments, 160 groups of experiment data were obtained, which were all divided into 100-s fragments for analysis, and 1920 valid time segments were obtained.

Data analysis

After the experiments, the data of recorder was imported into Tobii Studio software, and the scene in the driver’s field of vision was divided into nine non-overlapping fixation areas⁴² by trisecting in the horizontal and vertical directions, which were represented with area of interest (AOI). The location and coverage of each AOI are shown in Figure 7 and Table 5. Due to the limited space, the process of regional division is no longer detailed.

Figure 7.

Division of drivers fixation area.

Table 5.

Coverage of drivers fixation area.

AOI number	Coverage	AOI number	Coverage	AOI number	Coverage
1	Upper left area—road signs and facilities	4	Upper middle area—road signs and facilities	7	Upper right area—road signs and facilities
2	Left middle area—adjacent lane of left front	5	Central area—middle and distant road	8	Right middle area—right road and road facilities
3	Lower left area—left rearview mirror and left road	6	Lower middle area—road near the front and in-vehicle dashboard	9	Lower right area—right rearview mirror and right road

AOI: area of interest.

The variables and symbols of experiment acquisition and partial experiment data are shown in Tables 6 and 7, respectively.

Table 6.

Variables and symbols of experiment acquisition.

Variable	Symbol	Variable	Symbol
Emotion	Em	Visit duration (s)	$V_{D}$
Fixation count (n)	$F_{C}$	Mean value	Mean
Fixation duration (s)	$F_{D}$	Sum of all data	Sum

Table 7.

Part of experiment data.

Number	Em	$F_{C} (Sum)$	$F_{D} (Mean)$	$V_{D}$ (Sum)
1	Calm	70	0.66	66.02
	Anxiety	60	0.92	89.87
2	Calm	85	0.63	72.81
	Anxiety	61	0.88	89.23
…	…	…	…	…
n − 1	Calm	87	0.69	74.26
	Anxiety	66	0.96	85.00
n	Calm	85	0.63	66.64
	Anxiety	70	0.82	88.53

The analysis of drivers eye movement data such as fixation count, fixation duration, and visit duration in AOI is to obtain the difference of eye movement characteristics of female novice extroversion drivers under calm and anxiety by utilizing Tobii Studio and SPSS software.

Results

Fixation count

The spatial focus of driver’s attention is reflected over a period of time by fixation. The fixation count reflects the number of information that the observer needs to deal with in the process of visual search. To some extent, the interest degree of observer to the area was embodied by comparing the driver’s fixation count of a certain time under calm and anxiety. Larger fixation count and greater amount of information the brain processes indicate that the drivers pay more attention to AOI than others. In this article, the difference of drivers extraction information was obtained to evaluate the safety of driving through studying the fixation count distribution of female novice extroversion drivers in AOI under calm and anxiety. The distribution of fixation count is shown in Figure 8.

Figure 8.

Distribution of driver’s fixation count under calm and anxiety.

Paired sample t-test and analysis of variance are made on the above data using SPSS. The results are shown in Table 8. The difference in fixation count under the calm and anxiety of female novice extroversion drivers is significant ( $p < 0.05$ , p is the accompanying probability), the main effect of emotion is significant $(p < 0.05)$ , and the main effect of fixation area is significant $(p < 0.05)$ .

Table 8.

Paired sample t-test and analysis of variance on drivers’ fixation count.

Paired sample t-test						Analysis of variance
Emotion	Mean	Standard deviation	Correlation coefficient	t	p	Effect value	Test of inter-subjective effect
Emotion	Mean	Standard deviation	Correlation coefficient	t	p	Effect value	Emotion	Fixation area
Calm	8.44	6.336	0.968	1.933	0.000	F	5.452	68.742
Anxiety	7.00	7.730				p	0.045	0.014

According to the AOI group function of Tobii Studio software, a subject’s attention area to an image is showed by different colors combined with heat map, the times of vision dwell for a subject in a certain area is showed in Figure 9. In this article, heat map is created according to fixation count. Red represents the area with the most fixation points, followed by green, and there are many excessive levels among them.

Figure 9.

Heat map of driver’s fixation count under (a) calm and (b) anxiety.

The results showed that the distribution law of drivers’ fixation count in AOI was consistent under calm and anxiety as a whole. The fixation count in the middle AOI was higher than both sides. The drivers paid more attention to the left rearview mirror and dashboard than the other areas. Compared with the situation of anxiety, the search width of drivers in the horizontal direction was significantly higher and the fixation count was larger under calm. The drivers’ visual angle was narrowed under anxiety, which significantly affected the number of right fixation points. The attention bias toward the middle and left areas affected drivers’ comprehensive extraction of information.

Fixation duration

The fixation duration is the constant time of the visual axis center position, which is used to represent the time taken by the driver to extract valid information from the target being viewed. The difficulty degree of information being extracted and the driver’s interestingness in the target are reflected by the fixation duration. The variable can distinguish whether the driver has an attention bias in different road conditions. The distribution of fixation duration for female novice extroversion drivers on each AOI under calm and anxiety in this article is shown in Figure 10.

Figure 10.

Distribution of driver’s fixation duration under calm and anxiety.

Paired sample t-test and analysis of variance are made on the above data using SPSS. The results are shown in Table 9. The difference in fixation duration under the calm and anxiety of female novice extroversion drivers is significant $(p < 0.05)$ , the main effect of emotion is significant $(p < 0.05)$ , and the main effect of fixation area is not significant $(p > 0.05)$ .

Table 9.

Paired sample t-test and analysis of variance on drivers’ fixation duration.

Paired sample t-test						Analysis of variance
Emotion	Mean	Standard deviation	Correlation coefficient	t	p	Effect value	Test of inter-subjective effect
Emotion	Mean	Standard deviation	Correlation coefficient	t	p	Effect value	Emotion	Fixation area
Calm	0.66	0.174	0.688	−5.044	0.001	F	11.340	4.396
Anxiety	0.92	0.214				p	0.038	0.185

Using the AOI group function of the Tobii Studio software, the driver’s attention area is observed more intuitively combined with trajectory map. As shown in Figure 11, the difference of driver extraction information is obtained in order to evaluate the driving safety under two states. The visual interface of trajectory map can show the position of fixation point and the length of fixation duration (the size of dot) on static stimulus materials (such as a picture or scene).

Figure 11.

Trajectory map of driver’s fixation duration under (a) calm and (b) anxiety.

The results showed that the difference of drivers’ fixation duration was not significant in each AOI under calm, more concentrated on the middle area, tending to concentrate attention. Compared with the situation of calm, the fixation duration of drivers increased relatively under anxiety, and there was a significant difference in fixation duration for each AOI. For drivers, the fixation time of the middle and right areas became longer and the attention bias to the fixation target had occurred.

Visit duration

The visit duration is not the fixation duration. It is the time from the first fixation point appearing in AOI to the next fixation point moving out of AOI, that represents the total length which is the sum of each fixation duration and the saccade time between every two points of driver’s fixation time in a certain AOI. The visit duration is a comprehensive expression of the amount of information in an AOI and the difficulty degree of information processing, which is used to measure the driver’s comprehensive interest degree in different areas. The distribution of visit duration for female novice extroversion drivers in each AOI under calm and anxiety in this article is shown in Figure 12.

Figure 12.

Distribution of driver’s visit duration under calm and anxiety.

Paired sample t-test and analysis of variance are made on the above data using SPSS. The results are shown in Table 10. There is no significant difference in visit duration of female novice extroversion drivers under the calm and anxiety $(p > 0.05)$ , the main effect of emotion is not significant $(p > 0.05)$ , and the main effect of fixation area is significant $(p < 0.05)$ .

Table 10.

Paired sample t-test and analysis of variance on drivers’ visit duration.

Paired sample t-test						Analysis of variance
Emotion	Mean	Standard deviation	Correlation coefficient	t	P	Effect value	Test of inter-subjective effect
Emotion	Mean	Standard deviation	Correlation coefficient	t	P	Effect value	Emotion	Fixation area
Calm	7.34	5.97	0.932	−2.015	0.079	F	1.877	11.494
Anxiety	9.99	8.86				p	0.304	0.040

The results showed that the distribution law of drivers’ visit duration in AOI was consistent under calm and anxiety as a whole. Most of the drivers concentrated on the middle area, followed by the left area, but the visit duration for each AOI was different. It indicated that anxiety could cause drivers’ attention bias to traffic information, and there was no significant difference in overall visit duration. Under the situation of anxiety, the drivers got more traffic information through visit duration to middle area, which was less concerned about the left and right sides, and the attention bias appeared. Traffic accidents were more likely to occur due to ignoring important traffic information.

Discussion

Through the real-vehicle and virtual driving experiments in this article, the eye movement data of drivers under calm and anxiety was obtained, and the eye movement characteristics of female novice extroversion drivers in two emotion states were explored. The main conclusions were as follows:

Under calm and anxiety, the drivers fixation count in the middle AOI is higher than both sides, and the difference in fixation count is significant $(p < 0.05)$ . The search width in the drivers’ horizontal direction is significantly higher under calm, and the fixation count is larger. Under the situation of anxiety, the drivers’ visual angle is narrowed, which significantly affects the number of right fixation points.

The difference in fixation duration of drivers under calm and anxiety is significant $(p < 0.05)$ . Compared with the situation of calm, the fixation duration of drivers increases relatively under anxiety, and there is a significant difference in fixation duration for each AOI. The fixation time of drivers for the middle and right areas becomes longer, and the attention bias to the fixation target has occurred.

There is no significant difference in visit duration of drivers under anxiety $(p > 0.05)$ , and the main effect of the fixation area is significant $(p < 0.05)$ . Compared with the situation of calm, drivers’ visit duration exists more in middle areas and less in both sides under anxiety, anxiety can cause drivers attention bias to traffic information, and there is no significant difference in overall visit duration.

There are differences in drivers’ fixation of each AOI under anxiety. For drivers, more attention is focused on the middle area, the fixation count and visit duration on the left area are relatively more, and the fixation duration of right area is relatively long.

Eye movement characteristics of female novice extroversion drivers under calm and anxiety are explored in this article, which makes up for the lack of micro-study in the field of traffic safety and provides a basis for the study of driver emotions and intentions. However, there are still some deficiencies in this study: (1) due to human emotions are complex and diverse, only calm and anxiety during driving were studied in this article, and the analysis of eye movement characteristics under other emotions are the focus of follow-up research. (2) Anxiety is a dynamic, changeable, and complex psychological process, and the study of eye movement characteristics considering the dynamic evolution of driver emotions is needed to proceed. (3) Only representative eye movement characteristics of female novice extroversion drivers were analyzed in this article. More detailed drivers classification need to be further explored. (4) Some external factors such as driving environments and driver proficiency will bring a certain amount of error for the experiment results. The influence of external factors on the experiment results should be controlled in subsequent studies.

Conclusion

Anxiety is an important factor affecting driver’s behavior. It is not only a significant premise to determine the changing law of eye movement of female novice extroversion drivers under anxiety for identifying drivers’ intentions accurately and realizing active vehicle safety driving but also the important content for preventing accident. By analyzing a large number of experiment data of eye movement, we, respectively, contrasted the differences of fixation count, fixation duration, and visit duration of driver in AOI under calm and anxiety, which provided the theoretical basis for the study of driver’s visual behavior law and driving safety. From the perspective of “humanism,” the study of drivers’ eye movement characteristics under anxiety will contribute to the rational planning of transportation infrastructure and help to build and develop a traffic system with more social benefit. Taking into account the universal applicability, further research on driver’s eye movements and anxiety can focus on the differences in eye movements of drivers with different characteristics under anxiety or in different road scenes, in order to improve the driver’s risk-perception ability in traffic accidents. The theoretical basis of the auxiliary driving system of active vehicle safety is offered by the results.

Footnotes

Handling Editor: Tao Feng

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was supported by the Joint Laboratory for Internet of Vehicles, Ministry of Education-China Mobile Communications Corporation under the project No. ICV-KF2018-03, the State Key Laboratory of Automotive Safety and Energy (grant number KF16232), the National Natural Science Foundation of China (grant numbers 61074140, 61573009, 51508315, and 51608313), the Natural Science Foundation of Shandong Province (grant numbers ZR2014FM027 and ZR2016EL19), and the Project of Shandong Province Higher Educational Science and Technology Program (grant number J15LB07).

ORCID iD

Xiao-yuan Wang

References

Wang

Shan

et al . Prediction of bus travel time using random forests based on near neighbors. Comp Civil Infrastr Eng 2018; 33: 333–350.

Yao

et al . Transit network design based on travel time reliability. Transport Res Part C 2014; 43: 233–248.

. Traffic engineering. Beijing, China: China Communications Press.

Zhang

Research on visual cognition about sharp turn sign based on drivers eye movement characteristic. Int J Pattern Recogn Artif Intell 2017; 31: 1759012.

Gao

et al . Recognition method of construction conflict based on driver’s eye movement. Accid Anal Prev 2018; 113: 193–201.

Qin

Dong

et al . Influence of vehicle speed on the characteristics of driver’s eye movement at a highway tunnel entrance during day and night conditions: a pilot study. Int J Environ Res Public Health 2018; 15: 656.

Liang

Pan

et al . Influence of dynamic highway tunnel lighting environment on driving safety based on eye movement parameters of the driver. Tunnel Underground Space Tech 2017; 67: 52–60.

Decker

Stannard

Mcmanus

et al . The impact of billboards on driver visual behavior: a systematic literature review. J Crash Prev Injury Cont 2015; 16: 234–239.

Darja

Areh

Cvahte

Examination of driver detection of roadside traffic signs and advertisements using eye tracking. Transport Res Part F Psychol Behav 2016; 43: 212–224.

10.

Hudk

Madlek

The research of driver distraction by visual smog on selected road stretch in Slovakia. Procedia Eng 2017; 178: 472–479.

11.

Gung

MN.

Analysis on characteristics of drivers visual change on the roundabout. J Korea Entertain Ind Assoc 2016; 10: 293–300.

12.

Hong

Iwasaki

Furuichi

et al . Eye movement and driving behaviour in curved section passages of an urban motorway. Proc IMechE, Part D: J Automobile Engineering 2006; 220: 1319–1331.

13.

Zhao

Ding

Rong

et al . The effects of highway curves on driver gazing behavior in a driving simulator. Int Conf Chin Transport Profession 2011; 2011: 2336–2347.

14.

Shang

Zhu

Chunhua

et al . Impacts of grassland highway landscape elements on driver’s eye movement indices. China Safety Sci J 2017; 27: 19–24.

15.

Zhi-Hong

LI.

Modeling analyses on driver’s gazing behavior based on real vehicle eye movement experiment. Sci Tech Eng 2018; 18: 312–317.

16.

Dickins

DSE

Lipp

. Visual search for schematic emotional faces: angry faces are more than crosses. Cognition Emotion 2014; 28: 98–114.

17.

Kleberg

Selbing

Lundqvist

et al . Spontaneous eye movements and trait empathy predict vicarious learning of fear. Int J Psychophysiol 2015; 98: 577–583.

18.

Park

Lewis

Wang

et al . Are you mad at me? Social anxiety and early visual processing of anger and gaze among Asian American biculturals. Culture Brain 2018; 6: 151–170.

19.

Byrow

Peters

The influence of attention biases and adult attachment style on treatment outcome for adults with social anxiety disorder. J Affect Disord 2017; 217: 281–288.

20.

Holas

Krejtz

Cypryanska

et al . Orienting and maintenance of attention to threatening facial expressions in anxiety–an eye movement study. Psychiatr Res 2014; 220: 362–369.

21.

Allsop

Gray

Flying under pressure: effects of anxiety on attention and gaze behavior in aviation. J Appl Res Memory Cognition 2014; 3: 63–71.

22.

Pasupa

Chatkamjuncharoen

Wuttilertdeshar

et al . Using image features and eye tracking device to predict human emotions towards abstract images. Image Video Tech 2016; 4: 419–430.

23.

Zimasa

Jamson

Henson

Are happy drivers safer drivers? Evidence from hazard response times and eye tracking data. Transport Res Part F Traffic Psychol Behav 2017; 46: 14–23.

24.

El Haj

Nandrino

Antoine

et al . Eye movement during retrieval of emotional autobiographical memories. Acta Psychologica 2017; 174: 54–58.

25.

Schmidt-Daffy

Fear and anxiety while driving: differential impact of task demandsspeed and motivation. Transport Res Part F 2013; 16: 14C28.

26.

Yao

Chen

Cao

et al . Short-term traffic speed prediction for an urban corridor. Comp Civil Infrastruct Eng 2017; 32: 154–169.

27.

Taylor

Deane

Podd

JV.

Determining the focus of driving fears. J Anxiety Disord 2000; 14: 453–470.

28.

Greger

Norris

Thach

WT.

Spike firing in the lateral cerebellar cortex correlated with movement and motor parameters irrespective of the effector limb. J Neurophysiol 2004; 91: 576–582.

29.

Pourabdian

Azmoon

The relationship between trait anxiety and driving behavior with regard to self-reported Iranian accident involving drivers. Int J Prev Med 2013; 10: 1115–1121.

30.

Wester

Böcker

Volkerts

et al . Event-related potentials and secondary task performance during simulated driving. Accid Anal Prev 2008; 40: 1–7.

31.

Xie

et al . Analysis on traffic safety distance of considering the deceleration of the current vehicle. Intel Comput Tech Automat 2009; 3: 491–494.

32.

Panayiotou

Karekla

Leonidou

Coping through avoidance may explain gender disparities in anxiety. J Context Behavior Sci 2017; 6: 215–220.

33.

Asher

Asnaani

Aderka

IM.

Gender differences in social anxiety disorder: a review. Clin Psychol Rev 2017; 56: 1–12.

34.

Hosking

Liu

Bayly

The visual search patterns and hazard responses of experienced and inexperienced motorcycle riders. Accid Anal Prev 2010; 42: 196–202.

35.

Cestac

Paran

Delhomme

PT.

Young drivers sensation seeking, subjective norms, and perceived behavioral control and their roles in predicting speeding intention: how risk-taking motivations evolve with gender and driving experience. Safety Sci 2011; 49: 424–432.

36.

Hills

Thompson

Pake

MJ.

Detrimental effects of carryover of eye movement behaviour on hazard perception accuracy: effects of driver experience, difficulty of task, and hazardousness of road. Transport Res Part F Traffic Psychol Behav 2018; 58: 906–916.

37.

Yiqing

Xiaoyuan

Jin

A method for judging the driving tendency of automobiles based on decision tree model. J Shandong Univ Tech 2012; 26: 1–5.

38.

Meng

Ren

et al . The influence of driving experience on the physiological characteristics of the driver under stress scene. Int Conf Green Intell Transport Syst Safety 2018; 16: 529–536.

39.

Koegh

French

CC.

Test anxiety, evaluative stress, and susceptibility to distraction from threat. Eur J Personality 2001; 15: 123–141.

40.

Grupe

Nitschke

JB.

Uncertainty and anticipation in anxiety: an integrated neurobiological and psychological perspective. Nat Rev Neurosci 2013; 14: 488–501.

41.

Zung

WW.

A rating instrument for anxiety disorders. Psychosomatics 1971; 12: 371–379.

42.

Underwood

Crundall

Chapman

Selective searching while driving: the role of experience in hazard detection and general surveillance. Ergonomics 2002; 45: 1–12.