Research on the visual cognition patterns of exit guide sign viewing on freeway interchanges
Abstract
The development of freeway construction and the increasing coverage of the road network have led to increasing requirements for guide signs. This article investigated drivers’ visual cognition pattern regarding exit guide signs on freeway interchanges. A static visual cognition experiment with 32 participants was carried out. The route information volume (four levels) and destination information volume (seven levels) were selected as the variables. An eye-tracking system was utilized to record drivers’ eye movement indicators, such as eye movement time, saccade frequency, seek time, and fixation duration. The results indicated that the eye movement time, saccade frequency, and seek time are highly correlated with information volume and increase significantly with the increases in information volume; although the fixation duration has no correlation with information volume, the fixation duration value, saccade frequency, and seek time of destination information are significantly higher than those of route information, and the destination information fulfills a stronger guiding function during the driver’s trip. The corresponding threshold values of destination information are 5, 5, 4, and 3 under the four levels of route information, and the threshold value of route information is 3.
Introduction
Guide signs can convey necessary information regarding travel and route to drivers, enabling drivers to arrive at their destinations quickly and safely.1 According to national freeway crash statistics, approximately 10% of traffic crashes are related to road network construction, and more than 40% of these crashes occur in the exit area of the freeway.2 The exit guide sign, as an important part of freeway infrastructure, plays a role in traffic guidance and traffic control and directly affects traffic safety. In addition, as a pivotal connection between the driver and the road network, the content of guide signs can also influence the driver’s cognition. The reactions and decisions of a driver are directly related to traffic safety. With the rapid development of road network construction of the freeway system in China, many problems have been raised in the process of applying exit guide signs on freeway interchanges. The first problem is serious information overload. The increasing density of the road network, ineffective traffic management systems, and the concept of “more information is better than none” all contribute to information overload. The second problem is typesetting confusion; because of information overload, this phenomenon occurs frequently on freeway interchanges all over the country. The exit guide signs shown in Figure 1 are typical of freeway interchanges in China because of the lack of normalized special regulations; this situation imposes a burden on the visual cognition and psychological pressure of drivers and tends to cause risky driving.3
Several relevant regulations regarding freeway exit guide signs are specified in China National Standards of Road Traffic Signs (GB5768.2-2009); however, the exit guide signs of freeway interchanges are not specifically highlighted. Through several decades of research, the United States has created a relatively optimized system for guide signs. Clear rules and regulations regarding the content of guide signs are specified in the Manual on Uniform Traffic Control Devices4 and Standard Highway Signs (SHS). For example, Section 2E.10 Amount of Legend on Guide Signs in the Manual on Uniform Traffic Control Devices (MUTCD) specifies the following: No more than two destination names or street names should be displayed on any Advance Guide sign or Exit Direction sign. A city name and a street name on the same sign should be avoided. Where two or three signs are placed on the same supports, destinations or names should be limited to one per sign, or to a total of three in the display. Sign legends should not exceed three lines of copy, exclusive of the exit number and action or distance information. However, in the actual situation in China, two destination names or two street names cannot meet the demand for information; furthermore, with the cultural and geographical differences between the United States and China, certain rules and regulations in the MUTCD are not feasible in China.
During the last three decades, many scholars and experts have studied traffic signs from the perspective of visual cognition. The pioneering research can be traced back to 1989. Dingus et al.5 found that when drivers receive traffic sign information, they decrease the visual cognition time in each view and increase the view frequency. Miller6 proposed that in information processing the information volume of short-term memory is 7 ± 2 (blocks). The amount of information will affect the driver’s understanding and memorization of traffic signs, and the resulting memory load will affect the driving behavior.7 Therefore, the amount of information on guide signs should not be beyond humans’ cognitive capacity. Long and Kearns8 proposed methods to improve the effect of signs through a comparison experiment according to the reaction time based on the characteristics of the dynamic visual conditions. Compared with foreign studies, domestic research studies mainly focus on the information volume. In 2008, Du9 from Wuhan University of Technology analyzed the relationship between the traffic sign information volume and visual cognition time and compared the visual cognition ability to distinguish target route names and nontarget route names using eye tracking. He concluded that the cognition time of the target route name is significantly less than that of the nontarget route name and that the threshold value should not exceed 5. In 2009, Wang10 from Chang’an University concluded that the threshold value should be 6 based on a static sign visual cognition experiment.
The introduction of eye-tracking systems has greatly promoted research on guide signs, with most studies being conducted based on this system. In 2004, NHTSA11 of the United States conducted research that found that drivers’ eye fixation would move from the front lane when drivers were following cars. The duration for which their eyes moved away from the front lane when there was a car ahead was shorter than that without a car ahead. The duration of drivers’ gaze toward specific objects was typically in the range of 1.6–2.0 s. In 2006, Pan and Lin12 from Tongji University applied an eye-tracking system in the field of traffic safety and analyzed the effect of light conditions on the visual cognition distance for traffic signs under different driving speeds using an outdoor driving test.
Most research studies on traffic sign cognition were carried out near 2010. The rapid development of the road infrastructure in China has resulted in new requirements to study the application of sign information. Furthermore, several problems still remained to be solved. For example, there is no adequate research involving exit guide signs on freeway interchanges, and most related research has centered on guide signs in urban areas. In addition, most studies on the amount of information on guide signs have also focused on urban areas; few studies have focused on guide signs on freeway interchanges. Moreover, there is only destination information on the guide signs in urban areas. Thus, scholars have typically investigated the guide sign as a whole. Because both destination information and route information are provided on the exit guide signs at freeway interchanges, specific research on different parts of the guide sign must be conducted. Furthermore, many studies simply use the fixation duration without further analysis of the eye movement pattern, which is not appropriate. A more appropriate approach to studying eye movement patterns is the use of a more microscopic and comprehensive indicator system; therefore, four eye movement indicators were selected in this research.
Visual cognition is the driver’s most direct response to the sign and is a microscopic visual response rather than a physical action response. Based on visual cognition, this article aimed to obtain the cognition characteristics and patterns to guide signs on freeway interchanges. This article attempted to study the following:
1.
The visual cognition characteristics and patterns under the condition of different information volumes of guide signs on freeway interchanges;
2.
The different change patterns of eye movement indicators between destination information and route information;
3.
The threshold value of route information and destination information.
To achieve these objectives, a static visual cognition experiment was carried out using an eye-tracking system, which helped to obtain the eye movement data. Eye movement time (EMT), saccade frequency, seek time, and fixation duration were selected as the indicators to analyze the visual cognition patterns of different parts on the exit guide signs on freeway interchanges and to obtain the information threshold value of route information and destination information. The results of this study will hopefully improve the existing standards.
Method
Participants
A total of 32 healthy drivers, including 25 males and 7 females aged 20–60 years (mean = 33.93, standard deviation (SD) = 10.50), participated in this static visual cognition experiment. All participants had Chinese class C driver’s licenses with more than 3 years of driving experience (mean = 10.32, SD = 7.89) to ensure that they were adequately familiar with driving to identify the guide signs. All participants provided written informed consent before the experiment started.
Regarding the choice of the participants, briefly, the participants were a homogeneous sample selected to decrease the error caused by heterogeneity, as numerous studies have demonstrated that driving performance can be significantly affected by age and gender.13 Therefore, gender proportion of the participants in this article were selected according to the actual distribution of drivers in China. While the ratio of male to female drivers is about 3:1 in China,14 the numbers of drivers aged 20–30, 30–50, and over 50 are 14, 15, and 3, respectively. In addition, according to Central Limit Theory, if a sum of random variables is normally distributed, a large sample size obtained from those variables also fits normal distribution. Besides, the sample size not less than 30 is a rule of thumb and commonly used in driving behavior empirical research.15
Apparatus
The E-Prime software can be used to conduct behavioral and psychological experiments. In our experiments, images of guide signs were displayed randomly by E-Prime software, and then the software obtained the drivers’ response time in milliseconds. In this article, the response time data were not used; instead, they will be analyzed with the eye movement data in later studies.
The eye-tracking system (Figure 2) used in this article can realize the visual tracking of three-dimensional (3D) information, such as the real world, in a dynamic process. In the process of obtaining data, the eye-tracking system can improve not only tracking accuracy through its automatic deviation adjustment, but also portability through wireless real-time control and real-time recording. The parameters of the eye-tracking system are as follows: the sampling rate is 60 Hz; the tracking resolution is 0.1 degrees; the fixation accuracy is 0.5 degrees; the tracking range is 80 degrees in the horizontal direction and 60 degrees in the vertical direction; and the tracking distance is over 40 cm.
Experimental design and procedure
Division of signs
The layout of a freeway guide sign includes information on the destination, route, exit number, direction, and distance. This article divided the signs into four parts from top to bottom, as shown in Figure 3: exit number part, route information part, destination information part, and direction and distance part.
Each piece of information on the guide sign is critical because each provides a different type of signal for drivers. In this article, the amounts of route and destination information were selected as the experimental variables; compared with route information and destination information, the exit number and direction and distance information are fixed variables. Therefore, regardless of how much information is provided regarding route and destination, there is only one piece of information regarding exit number, distance, and direction. This situation implies that information overload is mainly caused by the overflow of route information and destination information. Thus, the route information part and destination information part were selected for investigation and analysis in this article.
Indicator selection
The basic parameters of eye movement directly selected by the eye-tracking system included fixation, saccade, and blink, as shown in Figure 3. Fixation points A, B, and C represented the place where the participants gazed, and the duration represented the length of time the participants gazed at that point. The saccade represented the eye movement from fixation A to fixation B, and the saccade frequency represented the duration of the eye movement. A blink represented an instance in which the viewpoint jumped suddenly from fixation B to fixation C.
The eye movement indicators derived from the basic parameters of eye movement in this article included EMT, saccade frequency, seek time, and fixation duration:
•
EMT refers to the time during which the participants identify the target information after the appearance of signs.
•
Saccade frequency is the number of saccades in the analysis section. A higher saccade frequency indicates that it is difficult for participants to find the target information.
•
Seek time is the time for the fixation point to enter the analysis area and finally locate the target information. Similar to saccade frequency, a longer seek time indicates that participants are taking longer time to find the target information.
•
Fixation duration indicates the time in which participants gaze at the target information; therefore, a longer fixation duration indicates that the participant is paying more attention to the information.
In addition, there is another indicator called visual recognition order. Through questionnaire survey and objective statistics, the subjective and objective visual recognition orders were obtained to determine the driver’s demand for different information.
Experimental factors
To reduce the deviation of experimental results due to familiarity, the names of the provincial capital city with two Chinese characters were selected as the target destination information, and the double-digit route number of the national freeway was selected as the route information (e.g. G20), as shown in Figure 3. The destination information (DI) volume was divided into seven levels (2-DI to 8-DI). The route information (RI) volume was divided into four levels (1-RI to 4-RI). Therefore, 28 combinations were considered in total. To reduce the inertia error, 32 signs were used as the interference factor in this study; thus, a total of 60 freeway guide signs were used in this experiment.
In the process of making experimental signs, the information was selected randomly, as was the target information for each sign. Before the experiment, the experimental signs were input into the E-Prime software, and the random screening mode was checked to ensure that each experiment sign would appear in a random order.
Experimental process
Relevant pictures regarding the experimental process are shown in Figure 4. The experiment was carried out with the following steps:
1.
Pre-experiment questionnaire. All participants are required to complete a questionnaire to collect basic information (e.g. age, profession, gender, and driving experience) and determine their physiological and psychological status before the test.
2.
Read the experiment guide. After filling in a questionnaire, the experimental guidelines are introduced to the participants: (1) for the formal test, the target information (including one piece of destination information and one piece of route information) appears on the screen; (2) please remember the target information and press the “Y” button to proceed to the next page, on which the target guide sign will be displayed; and (3) after finding the target information on the sign, press “Y” button rapidly to enter the evaluation page and complete the identification difficulty evaluation (the result would be used in later research).
3.
Equip the participant with the eye-tracking system and adjust it. Before the experiment, the participant was equipped with an eye-tracking system. The experiment assistant points three different points on screen. Then the participant’s eyes follow the three points successively. In this way, the participant was fitted with the eye-tracking system.
4.
Start the experiment. Each experiment task will last approximately 10 min.
5.
Complete the questionnaire survey after the experiment. After the experiment was completed, all participants were required to finish another questionnaire, which was designed to evaluate the difficulty of seeking target signs.
Analysis and results
Data preprocessing
The eye-tracking system helped obtain the video that contained the gaze point and the TXT format file that provided the basic parameters of eye movement. The data analysis was performed using BeGaze software, which is a software for an eye-tracking system that is used to analyze the data and video. The raw data of the route information part and destination information part were obtained under its function of selecting a specific part.
Analytical methods
The EMT, saccade frequency, seek time, and fixation duration were selected as the indicators in this study, and the route and destination parts were selected as the analysis parts. In this article, main effect analysis and analysis of variance with repeated measures (rANOVA) were used to analyze the eye movement indicators. Principal component analysis was selected to investigate the threshold value under the condition of multiple indexes. The analysis framework is shown in Figure 5.
The analysis is divided into five parts, as shown in Figure 5.
1.
The visual cognition patterns of the guide sign, EMT, and visual recognition order are the indicators used to analyze this part because both indicators are obtained through the date of the entire guide sign;
2.
The visual cognition patterns of the route information part, seek time, saccade frequency, and fixation duration of the route information part are the indicators used to analyze this part;
3.
The visual cognition patterns of the destination information part, seek time, saccade frequency, and fixation duration of the destination information part are the indicators used to analyze this part;
4.
The relationship between the route information part and the destination information part;
5.
The threshold value of information obtained from the level of recognition contains the threshold value of destination information and route information.
Analysis and results
The guide sign
The visual cognition patterns of the guide sign included two parts; the first part was the visual recognition order of route and destination information.
The research on visual cognition orders was carried out from subjective and objective perspectives, as shown in Figure 6. The subjective result was obtained through a questionnaire survey of 32 participants, and the objective data were obtained through the actual visual cognition to 28 target signs by 32 drivers (28 × 32 = 896 in total). The results showed that the proportion of priority recognition of destination information was slightly higher than that of route information, indicating that the destination information fulfills a stronger guiding function during the driver’s trip.
The second part of the visual cognition patterns of the guide sign was the EMT of different guide signs which contained different amounts of information.
The EMT results are shown in Figure 7(a)–(d). The results imply that the EMT increases as the amount of destination information increases. This trend is not linear but instead follows a quadratic equation (R2 > 0.85), indicating that the growth rate of the driver’s visual burden increases gradually with increasing destination information, and there is always an abrupt junction at which the EMT changes drastically. The abrupt junction indicates that the difference of EMT in neighboring destination information is the largest among the different destination information. The abrupt junction of 1-RI is 5-DI to 6-DI, as shown in the circle in Figure 7(a), and the abrupt junction of 2-RI, 3-RI, and 4-RI is 6-DI to 7-DI, as shown in the circles in Figure 7(b)–(d), respectively. The growth rate of EMT declines gradually with increasing route information, as shown in Figure 7(e). The linear growth slope of EMT is 0.1478, 0.1207, 0.1399, and 0.088 ((EMT8-DI– EMT2-DI)/6) under the condition of 1-RI, 2-RI, 3-RI, and 4-RI, respectively. After 6-DI, the growth rate of EMT was largely consistent, as shown in Figure 7(f).
According to the main effect analysis, there is a significant relationship between the EMT and destination information (p < 0.01), and there is no significant relationship between the EMT and route information (p = 0.122). Thus, the information was grouped into different levels by the S-N-K: the destination information was grouped into four levels (2-DI & 3-DI; 4-DI & 5-DI; 6-DI; 7-DI & 8-DI), and all route information was grouped into one level (1-RI, 2-RI, 3-RI, & 4-RI). The information levels stand for the different EMT; on the other hand, the EMT value of the high information volume group was significantly higher than that of the low information volume group; this difference can help observe the characteristics and patterns of visual cognition.
The EMT is jointly determined by the destination information and route information; therefore, to further analyze the influence of the guide signs, the influence parts of the destination and route information should be separated. However, there is an interaction effect between the destination and route information (p < 0.001); therefore, the influence of route information should be considered when analyzing the destination information area. Similarly, the influence of destination information should be considered when analyzing the route information area.
Route information part
As illustrated in Figure 8(a) and (b), there is an increasing trend in saccade frequency and seek time with increasing route information. In addition, there is a smaller gap between 2-RI and 3-RI in terms of saccade frequency and seek time; this observation indicates that a higher amount of route information makes it more difficult for participants to find the target information. In terms of fixation duration, there is no significant difference among the four levels of route information within each DI group, as shown in Figure 8(c), which indicates that the fixation duration of route information is not related to the information volume.
According to the results of the statistical analysis, the saccade frequency and seek time had significant relationships with the amount of route information. Moreover, the saccade frequency was divided into three groups by the method of S-N-K (1-RI; 2-RI & 3-RI; 4-RI), and the seek time was divided into four groups according to four levels of route information. In contrast, the fixation duration was not significantly influenced by route information (p = 0.213) because its range was 0.2–0.3 s, which indicates that the fixation duration will not change as the amount of information increases, as shown in Figure 8(c).
In conclusion, an increase in the amount of route information will increase the difficulty of finding the target information but will not affect the driver’s fixation duration on the target information.
Destination information part
Similar to the route part, there is a growing trend in saccade frequency and seek time with an increasing amount of destination information, as shown in Figure 9(a) and (b). Regarding the fixation duration, all fixation duration points of route information fall within 0.4–0.5 s, which is higher than that of the route part, and there is no clear fluctuation in the destination information, as shown in Figure 9(c). In addition, the eye movement parameters in the destination part do not change considerably with increasing route information.
Saccade frequency and seek time exhibit highly significant relationships with destination information (p < 0.01 for both). Furthermore, the saccade frequency was divided into five groups by the method of S-N-K (2-DI; 3-DI & 4-DI; 5-DI; 6-DI; 7-DI&8-DI), and the seek time was divided into six groups (2-DI; 3-DI & 4-DI; 5-DI; 6-DI; 7-DI; 8-DI). In contrast, the fixation duration is significantly influenced by the amount of route information (p = 0.061).
Relationship between route information and destination information
The destination information and route information play different roles in the guide signs, that is, they convey different messages to a driver and show different patterns in the eye movement parameters. The eye movement parameters of route information and destination information with the same amount of information are shown in Figure 10. The saccade frequency and seek time have similar trends with the two types of information. Under an information volume of 2, route information and destination information yield similar saccade frequencies and seek times. With increasing information, the gap between the saccade frequencies and seek times of route information and destination information increases when the information volume is 3 and then decreases when the information volume is 4, as shown in Figure 10(a) and (b). The fixation duration is higher for the destination information than for the route information.
Under the same amount of information, the eye movement parameters of two types of information were selected to measure participants’ sensitivity to different types of information; the analysis results are shown in Table 1.
| Indicator | Average | SD | p-value | ||
|---|---|---|---|---|---|
| RI | DI | RI | DI | ||
| Saccade frequency | 3.105 | 3.375 | 0.518 | 0.849 | 0.010 |
| Seek time | 0.324 | 0.380 | 0.099 | 0.141 | 0.002 |
| Duration fixation | 0.292 | 0.407 | 0.072 | 0.086 | 0.000 |
SD: standard deviation; RI: route information; DI: destination information.
As displayed in Table 1, all destination information values are higher than the route information values in terms of saccade frequency, seek time, and duration fixation, and the p-values of saccade frequency, seek time, and duration fixation are 0.010, 0.002, and 0.000, respectively. This result indicates that under the same amount of information it will take more time for participants to seek and stare at the destination information than at the route information. Moreover, the SD of destination information is higher than that of route information in terms of saccade frequency, seek time, and duration fixation, indicating that drivers are more volatile when they identify destination information.
Threshold value of information
Information, including the route information and destination information, is one of the main study objectives of this article; therefore, in the study of information threshold, this article will also consider the road name information and place name information. To determine the threshold value of destination information, principal component analysis was used in this study to obtain composite indicators. The original indicators were the EMT (VD1), the saccade frequency of the destination information part (VD2), the seek time of the destination information part (VD3), and the fixation duration of the destination information part (VD4).
Taking 1-RI as an example, the Kaiser–Meyer–Olkin (KMO) test statistics of four original indicators was 0.795, which was higher than 0.5. The statistical value of the Bartlett spherical test was 30.441 (p < 0.001). After the raw data were normalized by Z, the correlation coefficient matrix table of each index was obtained. The absolute values of the data in Table 2 were above 0.3, indicating a strong correlation between the variables.
| Indicator | VD1 | VD2 | VD3 | VD4 |
|---|---|---|---|---|
| VD1 | 1.000 | |||
| VD2 | 0.919 | 1.000 | ||
| VD3 | 0.942 | 0.990 | 1.000 | |
| VD4 | 0.901 | 0.876 | 0.895 | 1.000 |
Based on the principal component analysis, each indicator’s eigenvalue and contribution rate were calculated, and the total variance of the interpretation was obtained, as shown in Table 3. According to the size of the eigenvalues (generally greater than 1), two factors (F1, F2) were selected. The extracted information accounted for 97.795% of the raw data information.
| Component | Initial eigenvalue | Extract the sum of squares and loads | ||||
|---|---|---|---|---|---|---|
| Total | Variance % | Accumulated % | Total | Variance % | Accumulated % | |
| 1 | 3.762 | 94.053 | 94.053 | 3.762 | 94.053 | 94.053 |
| 2 | 0.150 | 3.742 | 97.795 | 0.150 | 3.742 | 97.795 |
| 3 | 0.080 | 2.008 | 99.803 | |||
| 4 | 0.008 | 0.197 | 100.000 | |||
The extraction method used was principal component analysis.
Through the regression algorithm, the loading formulas of two common factors were obtained
(1)
(2)
According to formulas (1) and (2) and the proportion of the corresponding eigenvalues of two common factors in the total extracted eigenvalues, the comprehensive scoring calculation method was obtained as follows
(3)
The comprehensive score of different traffic information under the condition 1-RI was calculated by formula (3). Similarly, the comprehensive scores of different traffic information under the conditions of 2-RI, 3-RI, and 4-RI were also calculated. The comprehensive score represents the recognition burden (RB) in a certain sense: a higher score represents that the driver bears a heavier burden. Different DIs correspond to the RB, and the relative change rate (RCR) of the RB between adjacent information represents the trend in the RB. The relationship between RB and RCR is shown in the following formula
(4)
where i and j represent the destination information volume, and both are less than or equal to 8.
In this article, the RCR was used to measure the information threshold, as shown in Table 4.
| DI | 1-RI | 2-RI | 3-RI | 4-RI | ||||
|---|---|---|---|---|---|---|---|---|
| RB | RCR | RB | RCR | RB | RCR | RB | RCR | |
| 2-DI | −0.97586 | − | −0.79257 | − | −0.89305 | − | −1.00826 | − |
| 3-DI | −0.76235 | 0.21351 | −0.66795 | 0.12462 | −0.76801 | 0.12504 | −0.78248 | 0.22578 |
| 4-DI | −0.74825 | 0.01410 | −0.54510 | 0.12285 | −0.62224 | 0.14577 | 0.17604 | 0.95852 |
| 5-DI | −0.39145 | 0.35680 | −0.46493 | 0.08017 | 0.22136 | 0.8436 | 0.68496 | 0.50892 |
| 6-DI | 0.549381 | 0.94083 | 0.65262 | 1.11755 | 0.52345 | 0.30209 | 0.95860 | 0.27364 |
| 7-DI | 1.008724 | 0.45934 | 1.05229 | 0.39967 | 0.93454 | 0.41109 | 1.13925 | 0.18065 |
| 8-DI | 1.31981 | 0.31109 | 1.47089 | 0.4186 | 1.04669 | 0.11215 | 1.10117 | −0.03808 |
DI: destination information; RI: route information; RB: recognition burden; RCR: relative change rate.
There was a difference in the comprehensive scores of the seven levels of destination information. The lowest comprehensive score was 2-DI, while the highest comprehensive score was 8-DI, and the maximum RCR was between adjacent information 5-DI and 6-DI. The result of comprehensive evaluation is consistent with the cognition pattern to a certain degree, highlighting the objectivity and rationality of the method. Therefore, the threshold value of the destination information should be 5 under the condition of 1-RI. Similarly, the threshold values of destination information are 5, 4, and 3 under the conditions of 2-RI, 3-RI, and 4-RI, respectively.
According to the same method, EMT (VR1), the saccade frequency of the route information part (VR2), the seek time of the route information part (VR3), and the fixation duration of the route information part (VR4) were selected as the original indicators.
The KMO test statistic of the four original indicators was 0.665, and the statistical value of the Bartlett spherical test was 59.772 (p < 0.001). The extracted information comprised 85.454% of the raw data information. The results (RB) are shown in Table 5.
| RB of 1-RI | RB of 2-RI | RB of 3-RI | RB of 4-RI | |
|---|---|---|---|---|
| 2-DI | −1.37498 | −0.40609 | −0.4103 | 0.54594 |
| 3-DI | −1.37533 | −0.30717 | −0.3367 | 0.6354 |
| 4-DI | −1.32032 | −0.43438 | −0.16925 | 0.63655 |
| 5-DI | −1.0684 | −0.43941 | 0.01318 | 0.97734 |
| 6-DI | −0.56027 | −0.27287 | 0.42029 | 1.01521 |
| 7-DI | −0.48309 | 0.46881 | 0.53842 | 1.32311 |
| 8-DI | −0.3328 | 0.61572 | 0.67525 | 1.46085 |
| Average | −0.93074 | −0.11077 | 0.10441 | 0.94206 |
| RCR | − | 0.81997 | 0.21518 | 0.83764 |
RB: recognition burden; RI: route information; DI: destination information; RCR: relative change rate.
The comprehensive scores increase with an increasing amount of route information and destination information, and the average RBs are −0.93074, −0.11077, 0.10441, and 0.94206 under the conditions of 1-RI, 2-RI, 3-RI, and 4-RI, respectively. Analysis of variance was selected to analyze the RB of information of the four routes, and the results showed that the RBs of 2-RI, 3-RI, and 4-RI were higher than that of 1-RI, the RB of 4-RI was higher than those of 2-RI and 3-RI, and the RBs of 3-RI and 4-RI had no significant relationship. Moreover, the maximum RCR was between adjacent information 3-RI and 4-RI; therefore, the threshold value of the route information should be 3.
Discussion
The objective of this article was to observe the pattern of drivers’ visual cognition regarding route information and destination information on guide signs on freeway interchanges and to investigate the threshold value of destination information. A static visual cognition experiment was carried out using an eye-tracking system to obtain participants’ eye movement parameters. The main factors in this article were the amounts of destination information and route information. In addition to external conditions, the layout of information, such as the font, size, and color of the sign, can also have a large effect on the visual cognition of the sign and further affect the information threshold value; however, these factors were not studied in this article. The experimental sign was made directly based on the requirements of the China National Standards of Road Traffic Signs (GB5768.2-2009).
The freeway connected by interchanges can be divided into direct routes and indirect routes. Drivers can take a direct route through a single freeway interchange and take an indirect route through more than two freeway interchanges. In general, there is only one direct route marked on the exit guide signs on freeway interchanges, and all other route information involves indirect routes. To represent the indirect route, the United States adds “to” before the route information; the United Kingdom adopts the bracket form; Sweden adopts a virtual frame form; and France uses different colors. In contrast, no relevant regulation on indirect routes exists in China. Therefore, the indirect routes were not considered in this article. Future research will focus on the expression of the indirect route.
Visual cognition, decision, and reaction all belong to the process of driver’s feedback response to external stimuli. As the first and most direct step of this process, visual cognition has a high correlation with reaction, and the increasing trend with an increasing amount of information is highly similar between the EMT and reaction time; however, there are also differences. The threshold value of route information was 3 based on the visual cognition in this article, and the threshold values of the destination information were 5, 5, 4, and 3 under the different levels of 1-RI, 2-RI, 3-RI, and 4-RI, respectively. The difference was that the threshold value of route information was 4, and the threshold value of destination information was 6, 6, 5, and 5 successively by comprehensively considering the reaction time, EMT, and subjective evaluation obtained through the same experiment. The main reason for this difference may be that reaction is the last step of the process of driver’s feedback response to external stimuli and the reaction influenced by visual cognition. The information threshold in this article was determined by the visual data of the route and destination parts, which is stricter and more reasonable; therefore, relevant research should mainly consider the visual characteristics.
The static image viewing experiment limited the research result to some degree. To improve the results, the actual driving experiment would be conducted in future research. Besides, male and female in different ages were recruited in the experiment. Pradhan16 pointed out that there are significant age-related differences in driver scanning behavior. The impact of driver type on visual cognition would be studied in future works.
Conclusion
To solve problems existing in engineering applications, such as the information overload of freeway guide signs, a static visual cognition experiment was performed using an eye-tracking system that acquired participants’ eye movement parameters. The eye movement indicators acquired were the EMT, saccade frequency, seek time, and fixation duration. Principal component analysis and rANOVA were used to analyze the eye movement correlations. The four eye movement indicators and methods could also be used to analyze other signs. The following conclusions were drawn according to the research results:
1.
The EMT, saccade frequency, and seek time are highly correlated with information volume and increase significantly with an increasing amount of information. The growth of the EMT follows a quadratic equation, and the abrupt change point often occurs within the range of 5-DI to 7-DI. Each additional DI or RI will significantly increase the saccade frequency and seek time when drivers search for the target information. The fixation duration is not correlated with the amount of information.
2.
The proportion of priority recognition of destination information is slightly higher than that of route information, and the fixation duration value, saccade frequency, and seek time of destination information are significantly higher than those of route information, indicating that the destination information has a stronger guiding function during the driver’s trip and that drivers are more sensitive to destination information under a given amount of information.
3.
The RB and RCR were introduced by principal component analysis to determine the threshold value. According to the visual cognition pattern, the threshold values of the destination information are 5, 5, 4, and 3 under the four levels of 1-RI, 2-RI, 3-RI, and 4-RI, respectively. The threshold value of the route information is 3. The results could have a positive impact on the design of exit guide sign on freeway interchanges.
Visual cognition characteristics and patterns and their threshold values were investigated in this study. On the basis of this article, subsequent research studies will focus on the guide sign groups, the exit guide sign system of freeway interchanges, and indirect routes. In general, on a two-way, four-lane freeway, two guide signs are juxtaposed as a group: one denotes the information of the direct line and the other denotes the information of the exit line. On a two-way, six-lane freeway, three guide signs are installed as a group: two of them denote the information on the direct line and the other denotes the information on the exit line. Different combinations of guide signs should be applied to different types of freeway. Therefore, the information thresholds of different guide sign groups and their influences must be investigated.
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 National Natural Science Foundation of China (61672067) and Beijing Municipal Science and Technology Major Projects (Design and Demonstration of Landside Traffic Sign System for Beijing New Airport; Project No. D171100003917001).
ORCID iD
Zhanzhou Lin https://orcid.org/0000-0002-1589-1075
Footnote
Handling Editor: Jiangchen Li
References
1. He CM, Deng XD. Urban route “Language” research and practice of directional signage system. Beijing, China: China Architecture and Building Press, 2008.
2. Zhu ZB, Guo TY, Pan S. Accident’s severity prediction model for freeway exit ramp, vol. 9. Harbin: KEJI, 2018.
3. Liu B, Sun L, Jian R. Driver’s visual cognition behaviors of traffic signs based on eye movement parameters. J Transp Syst Eng Inf Technol 2011; 11: 22–27.
4. Federal Highway Administration. Manual on uniform traffic control devices, 2009, https://mutcd.fhwa.dot.gov/pdfs/2009/mutcd2009edition.pdf
5. Dingus TA, Hulse MC, Antin JF, et al. Attentional demand requirements of an automobile moving-map navigation system. Transp Res Part A 1989; 23: 301–315.
6. Miller GA. The magical number seven, plus or minus two: some limits on our capacity for processing information. Psychol Rev 1956; 63: 81.
7. Jamson SL, Tate FN, Jamson AH. Evaluating the effects of bilingual traffic signs on driver performance and safety. Ergonomics 2005; 48: 1734–1748.
8. Long G M, Kearns D F. Visibility of text and icon highway signs under dynamic viewing conditions. Human factors 1996; 38: 690–701.
9. Du ZG, Pan XD, Guo XB. Relationship between information quantity and visual cognition of traffic guide sign. J Traffic Transp Eng 2008; 8: 118–122.
10. Wang JJ, Wang J, Wu HG. A study on threshold value of traffic sign information overload. Highway 2009; 4: 174–180.
11. National Highway Traffic Safety Administration. Traffic safety facts: 2007 data: pedestrians. Ann Emerg Med 2009; 53: 824.
12. Pan XD, Lin Y, Guo XB, et al. Research on traffic sign discernible visual range under backlighting condition. J Highway Transp Res Dev 2006; 5: 29.
13. Calvi A, Benedetto A, De Blasiis MR. A driving simulator study of driver performance on deceleration lanes. Accident Anal Prev 2012; 45: 195–203.
14. Ge R, Huang K, Ying L. Research on restraint system: taking into female driver. J Chongqing Univ Technol 2017; 11: 16–24.
15. Li Y, Zhao X, He Q. Comprehensive evaluation and classification of interchange diagrammatic guide signs’ complexity. J Adv Transp 2018; 2018: 9865305.
16. Pradhan AK, Hammel KR, DeRamus R, et al. Using eye movements to evaluate effects of driver age on risk perception in a driving simulator. Hum Factors 2005; 47: 840–852.
Cite article
Cite article
Cite article
OR
Download to reference manager
If you have citation software installed, you can download article citation data to the citation manager of your choice
Information, rights and permissions
Information
Published In
Article first published online: March 5, 2019
Issue published: March 2019
Keywords
Rights and permissions
© The Author(s) 2019.
This article is distributed under the terms of the Creative Commons Attribution 4.0 License (http://www.creativecommons.org/licenses/by/4.0/) which permits any use, reproduction and distribution of the work without further permission provided the original work is attributed as specified on the SAGE and Open Access pages (https://us.sagepub.com/en-us/nam/open-access-at-sage).
Authors
Metrics and citations
Metrics
Journals metrics
This article was published in Advances in Mechanical Engineering.
VIEW ALL JOURNAL METRICSArticle usage*
Total views and downloads: 1266
*Article usage tracking started in December 2016
Altmetric
See the impact this article is making through the number of times it’s been read, and the Altmetric Score.
Learn more about the Altmetric Scores
Articles citing this one
Receive email alerts when this article is cited
Web of Science: 1 view articles Opens in new tab
Crossref: 1
- Utilization of Drivers’ Dynamic Visual Characteristics to Find the App...
Figures and tables
Figures & Media
Tables
View Options
View options
PDF/ePub
View PDF/ePubGet access
Access options
If you have access to journal content via a personal subscription, university, library, employer or society, select from the options below:
loading institutional access options
Alternatively, view purchase options below:
Purchase 24 hour online access to view and download content.
Access journal content via a DeepDyve subscription or find out more about this option.
