An Experiment on the Impact of Evacuation Signage Position on Wayfinding Efficiency in Subway Stations Based on VR Technology

Abstract

Subway stations are complex spaces with high passenger density and mobility, making evacuation efficiency particularly important. The position of evacuation signage is a key factor affecting the efficiency of passenger wayfinding. This study constructed a virtual subway station scenario by virtual reality technology (VR), recording and analyzing the evacuation time and exit selection of 60 participants. The results showed that hanging evacuation signage, with their superior visual advantages, can greatly improve passenger wayfinding efficiency, followed by signage affixed to walls and columns, and finally signage on the ground. This study has provided a theoretical basis for the scientific layout and optimization of evacuation signage positions in subway stations, thereby enhancing passenger safety and evacuation efficiency.

1. Introduction

As urban traffic congestion becomes increasingly prominent, more and more citizens are choosing to take the subway to save commuting time and avoid traffic jams. In Qingdao, by 2024, the public transportation share of subway use has reached a daily high of 61%. The rise in subway passenger traffic and the increased pedestrian density have not only imposed pressure on transport but also made it difficult for passengers to find their way around the complex subway. Therefore, considering wayfinding behaviors and solving wayfinding problems from a crowd perspective to ensure personnel safety is very important for passengers to complete their travel activities smoothly.

Passengers’ wayfinding behaviors are highly correlated with environment [1]. An easily recognizable spatial interface is crucial for the safe and efficient travel of passengers. Among the various elements of the spatial interface in subway stations, signage plays an important role. Signage, serving as a distinctive and fully visible directional reference in subway station, can directly improve the efficiency of wayfinding. D. Snopkova et al. have showed that wayfinding efficiency directly improved by signage can be viewed as the spatial behavior in which individuals convert environmental information into wayfinding decisions and action plans [2]. Reviewing previous related studies, most of the research perspectives focus on macro-level social population effects [3] and environmental complexity [4]. In the signage related field, most studies focus on factors such as color [5], shape [6], visibility [7], and size [8] of evacuation signage. These studies on signage information can indeed help passengers quickly determine their location to a certain extent, enabling them to follow instructions and swiftly choose the optimal evacuation route. However, it is worth noting that despite the abundance of previous research findings, few scholars have conducted in-depth quantitative assessments of passenger wayfinding efficiency from the perspective of the distribution of signage positions.

In the research process, adopting traditional field experiments not only interferes with normal subway station operations, but also faces multiple contingencies that do not compromise precise control of single variables. Research finding by M et al. have shown that Virtual Reality (VR) technology has the ability to build highly simulated experimental scenarios [9]. In such scenarios, the technology not only enables precise control of critical variables, but also simultaneous collection of reliable quantitative data [9]. Dong et al. also have employed this technology in related research, validating its science in experimental applications [10]. Based on these advantages, the researchers used virtual reality technology to simulate the subway stations realistically. In this way to break the physical-temporal limitations and safety risks. During the experiment, due to individual differences in the objective time wasted by participants during the evacuation process due to hesitation and decision-making, each participant’s evacuation time varied. At the same time, researchers collected participants’ subjective evaluations and conducted in-depth analysis and discussion of the experimental results based on these data.

This study has explored and constructed a research framework based on virtual reality technology to deeply analyze the influence of evacuation signage positions on wayfinding efficiency. It recorded detailed data on participants’ wayfinding times, exit choices and subjective feelings under different evacuation signage positions. The research results have not only provided precise guidance on how to scientifically set up evacuation signage in subway stations but also improved the efficiency of passenger wayfinding and ensured rapid evacuation of people in emergencies. By focusing on details, this study can help optimize the functionality of urban transportation hub spaces, minimize safety risks, and effectively ensure the safety of people’s lives.

2. Methodology

Researchers conducted a survey of subway stations in Qingdao, Shandong Province, systematically collected and categorized various types of spatial configurations. At the same time, the researchers comprehensively reviewed relevant standards for architectural design and fire safety and then used modeling software and VR technology to build an experimental scenario that closely matches the real environment. Based on statistical principles, researchers reasonably planned the experimental groups and formulated detailed and scientific experimental procedures. Participants were recruited extensively from among university students and guided to complete the experiment in an orderly manner. After the experiment, SPSS 27.0 was used to conduct validity testing and quantitative analysis on the collected raw data in order to thoroughly investigate the differentiated impact of evacuation signage positions on passengers’ wayfinding efficiency. The detailed methodological steps are shown in Figure 1.

This study has been approved by the academic committee of the author’s institution. All participants were informed of the research purpose, process, and potential risks of the study and signed a written informed consent form. The study was conducted in accordance with the requirements of the Declaration of Helsinki and relevant laws and regulations. The personal information and data provided by participants were known only to the researchers to ensure privacy protection.

The concourse level and platform level, as core areas with frequent changes in passenger flow behavior, are key areas of focus when planning the position of evacuation signage. In the daily operation of subway stations, the concourse level serves as a connection between the ground level and the platform level. These three spaces together constitute the scope of this study, as shown in Figure 2.

In the experiment, researchers listed the positions of evacuation signage in order from top to bottom according to visual habits. As indicated by the red box in Figure 3: on the ceiling, hanging, affixed to walls and columns, and placed on the ground. However, in real subway stations, in order to effectively suppress smoke diffusion in the event of a fire and buy more time for people to evacuate safely [11], smoke barriers are usually installed on the ceilings, as shown in Figure 4. Furthermore, the smoke generated by a fire can easily obscure evacuation routes and signage, causing people to become disorientated. In light of this, evacuation signages are not installed on the ceiling of subway station in actual operation, so this position is excluded.

2.1. Experimental Preparation

To ensure the scientific and comprehensive nature of the study, researchers consulted with subway design experts and conducted field investigations to select several subway stations in Qingdao that were complex and representative in terms of spatial structure and passenger flow routes. These subway stations are summarized into four types, as shown in Figure 5. Their spatial characteristics are shown in

Table 1. Characteristics of the four subway stations’ spatial layouts.

Type	Characteristic
Cross-shaped	The transfer distance is the shortest, but the phenomenon of passenger flow collision is likely to occur during peak hours.
T-shaped	It can flexibly switch between two-way transfers.
L-shaped	It is necessary to make a detour through the station concourse level, and the transfer time is relatively long.
Parallel-shaped	Platform-to-platform transfer can be achieved, but the arrival and departure times of trains need to be strictly controlled.

. Analyze and extract spatial characteristics and environmental information and use Sketch Up (2021) software to construct and design an accurate 1:1 scale model.

In the final model, the most representative “T-shaped” and “parallel-shaped” models were chosen. The platform level adopts island platform, with the concourse level set at a total length of 180 m and an effective width of 12 m. Elevators, escalators, and staircases were strategically placed between the platform level and concourse level, along with corresponding ticket gates and security screening machines, to realistically recreate the atmosphere of a subway station. To thoroughly eliminate the interference of spatial familiarity on experimental results [12], all platform and line names were modified, ensuring that the virtual environment remained entirely unfamiliar to participants. To further enhance the realism of the model, researchers applied realistic textures to the model in SketchUp (2021) based on on-site images. After the model was completed, the virtual environment was rendered using the Enscape 3.4.0 plugin.

To make the evacuation results more realistic, the model was specifically designed with three exits. During the experiment, as long as participants successfully evacuated through any of the exits, the evacuation was considered successful. In addition, to prevent participants’ evacuation routes from overlapping too much and to ensure smooth evacuation paths, three starting points were set up within the model. Since the objective of this study was to investigate the independent influence of evacuation signage’s position on wayfinding efficiency, the researchers prioritized controlling this core variable while treating spatial distance, number of turns, and scene complexity as background variables commonly found in real-world scenarios, without intervening in them excessively. Additionally, the complex variable of crowd density was excluded to prevent factors such as crowd congestion and following behavior from influencing the independent effect of signage position. Additionally, the frequency of use of the three starting points across different groups remained consistent in the experiment, ensuring that the influence of these background variables on wayfinding efficiency was homogeneous across groups and would not interfere with the core variable’s impact on the final results. The final model and the specific positions of the starting points are shown in Figure 6.

As shown in Figure 7, the final experiment was divided into four groups: Group-O (control group), Group-A (hanging), Group-B (on the wall column), and Group-C (on the ground). Considering that smoke may be generated due to a sudden fire when actual danger occurs, which may affect the visibility of evacuation signage [13], the researchers adjusted the brightness of the evacuation signage in the experiment to 50% of normal brightness, as referenced from Chen et al. Chen et al. also demonstrated that green and black are the most suitable colors for safety signage, as this combination is highly effective for evacuation signage, has low cognitive load, and is more easily visible [14]. Therefore, the researchers set the evacuation signage colors in this study to green and black. All evacuation signages are placed in accordance with the requirements of China’s Code for Fire Protection Design of Metro (GB 51298-2018) [15] and Code for Fire Safety of Buildings (Part 1: General Requirements, GB 13495.1-2015) [16], and International Organization for Standardization (ISO) standards. Both Chinese and international standards require that signages be placed in conspicuous positions that are not easily obstructed.

Then, the researchers selected typical signage styles corresponding to the height range in real-world scenarios and set the side length a of the square signage to 400 mm and the outer diameter d of the circular signage to 175 mm in accordance with the specifications. Their sizes and colors are shown in Figure 8. The rectangular and circular signages used in the final model are shown in Figure 9. The rectangular signage can be regarded as three square symbols joined together in accordance with the specifications. The regulations also stipulate that signage affixed to walls and columns in public areas on the platform level and concourse level must have their upper edges no more than 1 m above the ground, with a spacing of no more than 20 m, and no greater than the distance between two columns. Evacuation signage suspended from the ceiling must have their lower edges no less than 2.2 m above the ground, with their upper edges no less than 0.5 m below the ceiling surface. Signage on the ground must be spaced 2 m apart. In the final model, the upper edge of signage affixed to walls and columns is 0.9 m above the ground, with a maximum spacing of 15 m (not exceeding the distance between two columns). And the lower edge of evacuation signage suspended from the ceiling is 2.9 m above the ground, and the upper edge is 0.9 m below the ceiling surface. The spacing between signage on the ground in the model is 2 m. Meet Chinese and international standards.

During the experiment, Group O served as the control group, with all signage removed, including evacuation signages and the ‘safety exit’ text signage at the exit. In this way, a benchmark for wayfinding efficiency without signage intervention is established, and then the influence of signage position change is directly quantified, which is convenient for subsequent comparison and analysis with other three groups. Group A added evacuation signage at the hanging positions based on Group O. Group B added the evacuation signage at the wall and column positions to Group O. Group C added the evacuation signage at the ground positions to Group O. All four groups included the three starting points mentioned above. To avoid the influence of order effects and subjective factors on the experimental results, the experimental group and starting point for each participant were determined using a balanced Latin method and random sampling, thereby eliminating human and gender-related factors.

2.2. Participants

To reduce the interference caused by individual differences and ensure the reliability and comparability of the experimental results [17], the researchers decided to determine the scope of the experimental subjects as the group of university students. University student cohorts, engaged in similar learning process within relatively uniform educational systems, exhibit significant convergence in their knowledge accumulation and educational experiences. This homogeneity enables them to quickly understand the experimental rules and quickly engage in evacuation actions by assuming designated roles, which is conducive to the smooth progress of the experimental process and guarantees that the experiment will be completed on time.

In the preliminary phase of the experiment, 86 university students were recruited as participants through voluntary enrollment. During the screening process, gender factors were fully considered to ensure a 1:1 male-to-female ratio, and 60 participants were ultimately selected. The basic information of the participants is shown in

Table 2. Participants’ Basic Information.

Gender	Number	Mean Age /Year	Mean Height/m	Mean Eye Level/m	Mean Corrected Eyesight	Mean Eye Disorder	Eyestrain
Female	30	20	1.65	1.54	1.0	0	0
Male	30	21	1.73	1.62	1.0	0	0

. During the experiment, participants drew lots to determine which group they belonged to, with each group consisting of 15 people: Group-O (control group), Group-A (hanging), Group-B (on the wall column), and Group-C (on the ground). The final confirmed visual acuity of all participants was normal, with no visual impairments such as color blindness or eye diseases, and no excessive eye strain within one week prior to the start of the experiment. Before the experiment officially began, all participants were given a detailed explanation of the instruments used in the experiment and signed informed consent forms. After the experiment, participants received compensation in accordance with predetermined standards in recognition of their support and contribution to the research.

2.3. Laboratory Equipment Set Up

This experiment was conducted in the laboratory of the author’s institution from October to November 2024. As shown in Figure 10, we selected and arranged a laboratory with dimensions of 10 m (length) × 6.4 m (width) × 3.5 m (height). The laboratory is separated by a wall into two single rooms of different sizes, with the larger room serving as the formal laboratory and the smaller room as the pre-laboratory. The equipment and interior design in both rooms were identical to minimize the influence of other factors on the experimental results.

To optimize the experimental experience for participants, the study employed the PICO 4 All-in-One VR Headset. The device uses the Pancake optical solution to present a 4K+ ultra-visual screen, 105° large viewing angle and 90 Hz high refresh rate, and with the four-eye environment tracking and infrared optical positioning system, it achieves accurate 6-degree-of-freedom spatial positioning, which dramatically enhances the sense of immersion. Meanwhile, the device’s pupil distance adjustment has been upgraded to stepless mechanism for easier operation, and the Hyper Sense Vibration Handle provides realistic and dynamic vibration feedback, allowing participants to control their movement in the model via the joystick.

Regarding speed settings, we refer to relevant studies such as Lin et al., who set the speed of NPCs in virtual experiments to 0.7 m/s to 2.8 m/s [18], and Cosma et al. who set it to 1.2 m/s [19]. To avoid causing motion sickness in participants due to excessive speed, while also ensuring that the speed is realistic, this experiment ultimately selected 1.7 m/s as the constant movement speed after multiple trials and feedback from participants. Additionally, the virtual space required for the experiment was constructed using SketchUp (2021) and Enscape (3.4.0) software, with specific parameters detailed in

Table 3. Configuration of Hardware and Software Parameters.

	Equipment/Software	Parameters
Field Angle	PICO 4 All-in-One VR Headset	105°
Resolution		4320 × 2160, 1200 PPI
Velocity of Movement	SketchUp (2021) + Enscape(3.4.0)	1.7 m/s
Virtual Time		3:00 pm

.

2.4. Experimental Procedure

The specific experimental process is shown in Figure 11. Firstly, participants were brought into the pre-laboratory and asked to fill out the basic information questionnaire within the allotted time, as shown in

Table 4. Pre-Experiment Questionnaire Content.

Experiment Date:	Group:	Age:
Class, Name:	Contact Information:	Gender:
Binocular Vision (Including Correction):		Height:
Have you been using your eyes too much recently?

. They then sat for 3 min to relax and calm their minds before putting on the VR headset. This was followed by a pre-experimentation phase, in which participants learnt and adapted to navigating the virtual space, including moving forward, backward, turning, and ascending/descending stairs, familiarizing themselves with the virtual environment and ensuring that the experimental equipment operated smoothly. During this period, the researchers ask the participants how they feel and confirm that they are in good physical condition, with no symptoms of motion sickness such as dizziness or nausea. Afterwards, the device was removed, and researchers briefed them on the specific procedures and precautions for the formal experiment.

Throughout the experiment, participants only knew which group they belonged to and did not know whether they were in the experimental group or the control group. This design helped to better observe the experiment process and, when necessary, deal with unexpected situations that might arise in a timely and appropriate manner to ensure the safety of the participants. Four minutes before the official start of the experiment, participants were informed that if they felt unwell during the experiment, they could stop the experiment at any time.

After the formal experiment began, participants used the starting point selected on the platform level as their starting point in the virtual environment. Participants in the experimental group were required to follow the instructions and respond based on the existing evacuation signage information; participants in the control group made judgments based on their accumulated experience and intuition from daily life. All participants used joysticks to quickly make pathfinding decisions until they escaped to the ground floor level, at which point the experiment concluded. During the experiment, the visuals participants saw in the VR device were synchronized in real-time to a computer. Researchers accompanied participants throughout the process to ensure no interference and recorded evacuation times in real-time.

After the experiment, participants were asked to complete another questionnaire, including a subjective evaluation of their evacuation status and their overall impressions of the experiment. Responses were to be indicated by circling the corresponding option only. The specific content is shown in

Table 5. Post-Experiment Questionnaire Content.

Class, Name:		Starting Point (Fill1/2/3):
Whether you can quickly and easily identify the evacuation signage during the escape?
5 Very Easy	4 A Bit Easy	3 General	2 Not Too Easy	1 Basically Unrecognizable
Whether the evacuation signage you recognize are helpful in way-finding?
5 Very Easy	4 A Bit Easy	3 General	2 Not Too Easy	1 Basically Unrecognizable
Is there a misjudgment due to a misreading of the signage?
1 Yes		0 No
Did you relieve anxiety after seeing the evacuation signage?
1 Yes		0 No
Are you satisfied with this experiment?
5 Very Satisfied	4 Relatively Satisfied	3 General	2 Less Satisfied	1 Very Dissatisfied
Suggestions:		Exit (Fill1/2/3):

The experiment is over, thank you for your participation. Wish you a happy life!

. Group O only needed to answer the last question, while Groups A, B, and C needed to answer all questions.

3. Results

3.1. Data Collection

As shown in Figure 12 and

Table 6. Average Evacuation Time Form.

	Group O			Group A			Group B			Group C
Starting Point	1	2	3	1	2	3	1	2	3	1	2	3
Mean Value (t)/s	230	277	197	187	183	149	194	174	180	211	198	222
Mean Value (t)/s	235			173			183			210

, the researchers objectively recorded the evacuation time (t) of all participants from different Starting Points during the experiment and calculated the average evacuation time across different groups and Starting Points.

Table 7. Evacuation Exit Record Form.

	Group O			Group A			Group B			Group C
Starting Point	1	2	3	1	2	3	1	2	3	1	2	3
Exit	1	1	1	1	1	1	3	1	3	3	2	3
	1	3	1	1	3	3	3	1	3	3	3	3
	1	3	1	1	3	3	3	3	3	3	3	3
	2	3	1	1	3	3	3	3	3	3	3	3
	3	3	3	1	3	3	3	3	3	3	3	3

presents the Exit choices of individual participants. As shown in Figure 13 and Figure 14, researchers represented the subjective feelings of different participants by a radar chart. The results showed that, in terms of the average evacuation time values for the four groups, Group O > Group C > Group B > Group A. Among them, the difference between Group O and Group A was the most significant, with an average evacuation time difference of 62 s.

3.2. Data Validation

Researchers imported all evacuation time data collected from the experiment into SPSS 27.0 software. Firstly, the data distribution pattern was verified by the Shapiro–Wilk test, and it was confirmed that the four groups of data were all normally distributed, as shown in

Table 8. Normality Test for Participants’ Evacuation Time.

Group	Number	Mean	SD	Sig.
O	15	234.87	71.46	0.139
A	15	177.07	26.98	0.114
B	15	182.53	29.74	0.072
C	15	210.20	42.76	0.381

. Subsequently, a homogeneity of variance test was conducted. And the Levene’s test results indicated heteroscedasticity (p < 0.001). Nevertheless, the researchers proceeded with a one-way ANOVA analysis to determine whether there were statistically significant differences in the overall evacuation times across the four groups. The results of the one-way ANOVA analysis showed significant differences between groups (F = 5.020, p = 0.004). with an effect size of approximately 0.212, indicating a moderately large effect. Given the heteroscedasticity, the researchers used the Games-Howell test for multiple comparisons, supplemented by Tukey’s HSD test to further focus and refine the pairwise differences between groups. The results of the multiple comparisons showed that there was a significant difference in evacuation time between Group-O and Group-A (p = 0.041 < 0.05), with Group-O having a significantly longer evacuation time than Group-A. However, as shown in Figure 15, there were no significant differences between Group-O and Group-B, Group-O and Group-C, as well as between Group-A, Group-B, and Group-C, showed no significant differences. The aforementioned data analysis process and results provide statistical evidence for investigating the impact of subway signage position on evacuation time.

3.3. Data Analysis

In both Group O and A, the mean evacuation time from Starting Point 1 and 2 was longer than that from Starting Point 3. Most participants in these two groups choose to evacuate from Exit 1 when they departed from Starting Point 1, and most of them evacuated from Exit 3 when they departed from Starting Point 2. For Starting Point 3, participants in Group O mostly evacuated from Exit 1, while participants in Group A mostly chose Exit 3. On the contrary, for Group B and Group C, the average evacuation time was longer for both Starting Point 1 and 3 than for Starting Point 2. Moreover, participants in these two groups basically choose to evacuate through Exit 3, regardless of which Starting Point they departed from. Overall, the data trends of Group B and Group C were consistent, and the data trends of Group O and Group A were similar but different, and this difference was mainly reflected in the evacuation results guided by Starting Point 3.

As shown in Table 5, to objectively quantify the participants’ evacuation, they were arranged to complete a questionnaire at the end of the experiment. The questionnaire encompassed five dimensions: signage recognition rate, signage effectiveness rate, decision error rate, anxiety reduction level, and overall satisfaction with the experimental procedure. Three of the questions were rated on a five-point Likert scale from “5” to “1”. When the scoring was completed, the data for each group were statistically analyzed and the corresponding mean scores were calculated.

In the comparison of various data, since Group O only needed to answer the last question, the researchers did not analyze or explain the first four indicators for this group. In Group A, participants from Starting Point 1 gave a score of 4.4 for signage recognition, which was higher than the scores from the same Starting Point for Group B and Group C. When departing from Starting Point 2, participants in Group A gave a score of 4.2, which was also higher than Group B and Group C. When departing from Starting Point 3, participants in Group A gave a score of 4.6, which was still higher than that of Group B and Group C. After calculation, the average score for Group A was 4.4, which was significantly higher than that of Group B (3.9) Group C (3.9) and Group O (0). Overall, regardless of the Starting Point, participants in Group A scored higher on signage recognition rate than those in Group B, Group C, and Group O.

In Group A, participants from Starting Point 1 had a mean score of 4.6 for signage efficiency, which was lower than the corresponding scores for Group B and Group C. Participants from Starting Point 2 scored a mean of 4.8 for signage efficiency, which is higher than Group B and Group C. Participants from Starting Point 3 had a mean score of 5.0, which was also higher than the scores for Group B and Group C. In terms of the overall signage effectiveness score, Group A was 4.8, surpassing those of Group B (4.6), Group C (4.5), and Group O (0).

In Group A, participants departing from Point 1 gave score of 0.6 in decision error rate, which was lower than the corresponding scores in both Group B and Group C. Participants in Group A, who departed from Starting Point 2, scored the decision error rate as 0.4, again lower than Group B and Group C. Participants in Group A from Starting Point 3 scored the decision error rate at only 0.2, which was lower than those of Group B and C. The average error rate score in Group A was 0.4, lower than those in Group B (0.9) and Group C (0.8).

When collecting data related to anxiety relief, it was found that except for participants in Group B who started from point 2, who scored 0.8 points, and participants in Group C who started from point 1, who scored 0.6 points, the remaining participants in Groups A, B, and C all scored 1 point. Specifically, the average anxiety reduction level score for Group A was 1.0, and both Group B and Group C were 0.9, which shows that the score for Group A was higher than that of Group B and Group C.

Regarding experimental satisfaction scores, participants in Group O scored 4.8 points, while both Group A and Group C participants gave 4.9 points, and participants in Group B gave 4.7 points.

4. Discussion

4.1. The Relationship Between Evacuation Signage Positions and Wayfinding Efficiency

In the research domain of evacuation signage in public buildings, there have been many studies that clearly indicate that the performance of evacuation signage is a key factor affecting the efficiency of personnel evacuation [20], a viewpoint that is highly compatible with the importance of the conspicuous Position of evacuation signage, which is highlighted in this study. However, as a special type of public building, the spatial structure of subway stations is unique and differs significantly from that of ordinary public buildings.

This study used VR technology to simulate a subway station scenario and conducted experiments using evacuation time as the core metric for measuring wayfinding efficiency. It investigated the impact of the positions of evacuation signage within subway stations on passengers’ wayfinding efficiency. If the evacuation efficiency of the control group is taken as the benchmark unit 1, the evacuation efficiency of the other three groups all showed varying degrees of improvement.

Table 9. Characteristics and Optimization Effectiveness of Signage at Different Positions.

Position	Optimization Effectiveness	Characteristics
Hanging	26.38%	High-visibility continuous guidance information
On the Wall Columns	22.13%	Information Acquisition Delay
On the Ground	10.64%	attention deficits and visual processing limitations

summarizes the signage characteristics and optimization efficiency under different positions. When evacuation signage was hung beneath the ceiling, participants exhibited a 5.46% reduction in mean evacuation time compared to wall-columned positions, and a 17.62% reduction relative to grounded positions. Concurrently, wayfinding decision error rate decreased by 55.56% and 50%, respectively, under these conditions. This finding indicates that in complex public spaces such as subway stations, the strategic positions of signage play a role in enabling passengers to rapidly acquire critical information and make accurate decisions [21]. When evacuation signage is positioned at high and visible positions within passengers’ visual fields, the average evacuation time to reach Exits decreases significantly compared to placements on surrounding walls and columns or floors, while wayfinding success rates is also significantly improved.

In the experiment, participants consistently selected the same Exit when departing from both Starting Point 1 and 2 under conditions of no signage and only hanging signage. This suggests that the convergence between daily life experiences accumulated by individuals and signage information acquired from conspicuous visual fields leads passengers to select escape routes analogous to those under optimal signage positions, even in the absence of explicit guidance. Previous studies have indicated that variations in evacuation time are primarily attributed to three key behavioral factors: obstacle avoidance [22], retrograde movement [23], and conformity [24]. Although these behaviors are fundamentally coordinated movements for collective interests, individual actions such as wayfinding hesitation [25], decision-making pauses [26], and path retracing [27] during evacuation can significantly prolong evacuation time. As the proportion of these behaviors increases, participants’ movement distance rises, leading to prolonged evacuation time and elevating the cost of reconfiguring the virtual space for participants.

However, when participants started from Starting Point 3 without any signage or only hanging signage, their choice of evacuation Exits was completely different. Looking back at Figure 4, we can see that when participants started from Starting Point 1 and 2, they were in a transfer space and faced a lot of information and multiple choices, whereas when they started from Starting Point 3, they were not affected by the transfer space. Therefore, the reason why these two groups started from the same Starting Point but chose different evacuation Exits is that:

(1)

In scenarios without signage prompts, participants required time to familiarize themselves with the environment and construct a cognitive map of the virtual space, leading them to choose evacuation directions requiring minimal cognitive judgment, with the majority ultimately exiting through Exit 1;

(2)

When there were hanging signage, participants leveraged acquired spatial information to select optimal evacuation paths with minimized time expenditure, resulting in predominant utilization of Exit 3.

When participants were confronted with evacuation signage positioned on surrounding wall columns and on the ground, participants basically chose the way close to Exit 3, regardless of their Starting Points. After observing the evacuation process of the participants, the researchers found that during the process, the majority of participants opted to Exit the platform level through the staircase in the middle of Starting Points 2 and 3. When they reached the station concourse level from these two staircases, evacuation signage installed on surrounding walls, columns, and on the floor pointed to the nearest Exit 3. Among these two groups of participants, only a very small number of people indicated in post-experiment discussions that they had used their experience to evacuate for a short time after reaching the platform level, and only then noticed the evacuation signage, ultimately choosing evacuation Exit 1. This phenomenon fits with the results of the questionnaire filled out at the end of the experiment. When the evacuation signage was positioned at wall columns and ground level, they were positioned below the optimal view of the person, and the participants perceived the signage in these positions to a lesser extent compared to the hanging signage. During the evacuation process involving all participants, very few chose to evacuate through Exit 2. This was due to the presence of the more visible Exit 1 next to Exit 2. During evacuation, participants typically rush toward the first identified Exit immediately upon discovery. However, only a minimal number of participants perceived Exit 2 and opted to evacuate through this alternative route.

However, this study still has certain limitations. In terms of the selection of dependent variables, using only evacuation time and exit selection as measurement indicators makes it difficult to fully reveal the underlying mechanisms by which signage position affects wayfinding efficiency. Future research could further explore: the economic efficiency of wayfinding behavior through path length analysis, the extent of participants’ attention to signage through gaze duration quantification, and the interference of signage on decision stability through the number of decision hesitations. This would enable a more comprehensive analysis of the causal chain effect of ‘signage position-attention allocation-path selection-evacuation time’, thereby elucidating the underlying mechanisms through which signage position influences wayfinding efficiency.

In addition, when determining the scope of the preliminary sample, in order to make the differences between the experimental group and the control group more likely to be attributed to the core variable of signage position and to reduce the interference of individual differences on the experimental effect, researchers selected university students, one of the high-frequency user groups of subway stations, as the subjects of the experiment. However, the actual user group of subway stations is diverse, covering people of different ages and physical conditions, and their perception and response to evacuation signages vary significantly: older people may rely more on large font signages, children may be more sensitive to colorful signages, and visually impaired people may need tactile or auditory auxiliary signages. Therefore, the conclusion drawn in this study that hanging signage produces the optimal wayfinding effect may not be directly applicable to the aforementioned special groups. Researchers hope to include children, the elderly, and visually impaired individuals in future studies, comparing and analyzing the preferences for signage positions and differences in wayfinding efficiency among different groups, and designing signage schemes tailored to specific groups to enhance the universality and practical value of the research conclusions.

4.2. Recommendations for Subway Station Signage Design

A detailed analysis of the questionnaires completed by participants further revealed the key role of Signage position in wayfinding. The data shows that signage hung at a high level have a significant advantage in terms of visibility. Compared with the hanging signage, the signage at the wall columns is often obscured by the crowd or part of the obstacles, and their visibility is greatly reduced. However, participants’ attention during the search process is focused more on the path ahead and changes in the surrounding environment, making it difficult to notice ground signage immediately. The hanging signage, on the other hand, by virtue of their position above the crowd, were able to easily enter the field of vision of the participants and were the first to be captured.

However, the effectiveness of signage may also be influenced by crowd density at subway stations. Although this study did not directly include crowd density as a variable, it is possible to infer the impact of crowd density on different types of signage by considering real-world subway station operational scenarios. During peak hours when crowds are dense, passengers’ vision may be obstructed by people in front of them. In such cases, hanging signage at higher positions may have an advantage, while signage at lower positions may become ineffective due to being obstructed by the crowd, especially when passengers are caught up in the flow of people, as the likelihood of looking down at the ground or closely observing walls is significantly reduced.

From the perspective of guiding participants’ evacuation behaviors, hanging signage also performed better in terms of wayfinding efficiency. During the tense evacuation scenarios, participants are often required to make rapid action decisions in a very short period of time. The clear and explicit guidance information provided by hanging signage enables participants to rapidly identify evacuation directions, significantly reducing decision error rates during wayfinding. In comparison, some of the signage at the wall columns is relatively hidden and there is a delay in obtaining information; and the ground signage is affected by visual limitations, making it difficult to provide timely and effective guidance to participants during emergencies.

Hanging evacuation signage also plays a significant role in alleviating participants’ evacuation anxiety. Psychological studies show that timely access to information and clear guidance during crisis situations can effectively alleviate individual anxiety levels [28]. The good visibility and guidance of hanging signage provide participants with clear hanging signage during evacuation scenarios, thus obtaining psychological security and significantly relieving their anxiety. However, due to the lack of visibility and guidance of wall and column and grounded signage, participants were required to expend greater cognitive effort in searching for effective information, and uncertainty increases, which in turn aggravates anxiety. Therefore, hanging signage shows clear advantages in visual perception, wayfinding decision-making, and psychological comfort during evacuation processes compared to wall-columned and grounded signage.

Based on the above research results, researchers propose that the practical design of subway station wayfinding systems should prioritize the strategic layout of hanging signage. For example, moderately increase the number of hanging signage and rationally reduce the spacing between signage. With the advantage of a good field of vision, it can effectively avoid being obscured by the crowd, thereby maintaining good visibility in complex passenger flow environments. In areas where the guidance efficiency of hanging signage diminishes, a systematic and comprehensive installation of wall- columned and grounded signage should be installed to relay guidance, thereby forming a multi-dimensional and complementary guidance system, such as Figure 16. It is noteworthy that the specific setting of the hanging signage will need to be dynamically adjusted according to the spatial layout of the subway station and the real-time passenger flow [29,30]. There are significant differences between different subway stations in terms of functional positioning and passenger flow characteristics. Therefore, it is necessary to combine the surrounding facilities of the station to accurately determine the focus of passengers’ attention, and to design and layout hanging signage in a targeted manner to ensure that passengers can quickly capture key information.

Although hanging signage performs exceptionally well across various metrics, its practical application still requires consideration of multiple limiting factors. In reality, there are numerous functional signages within subway stations, such as transfer guidance signages and exit information signages. If the signage at the hanging position overlaps with these signages in terms of visual focus, it may lead to information overload, increasing passengers’ cognitive load and consequently reducing wayfinding efficiency. At the same time, the universality of signage design and layout must be assessed in conjunction with specific contextual factors. Some older subway stations are constrained by architectural limitations, such as insufficient ceiling height, which may prevent the installation of hanging signage and hinder the formation of an effective guidance system. This study only analyses wayfinding efficiency from a single perspective and does not take into account engineering factors such as the cost and durability of the identification system. In practical applications, the optimal choice of wayfinding scheme needs to be based on multi-objective decision-making, such as efficiency, cost, compatibility, etc., in order to finally determine the personalized scheme suitable for specific stations.

In addition, the illuminance of evacuation signage is an equally important factor in evacuation efficiency. Recent scholars have conducted quantitative studies by adjusting the luminance in experiments, showing that when the luminance is increased, it is easier for participants to recognize signage and make correct evacuation decisions [31]. This research provides novel empirical evidence highlighting the role of environmental factors in evacuation processes, while proposing innovative optimization strategies for subway station evacuation signage systems through spatial cognition-driven design principles.

Due to variations in cultural backgrounds and cognitive habits among populations in different regions, the design elements and information transmission methods of evacuation signage vary from place to place. These differences will in turn affect the visibility and accessibility of evacuation signage to passengers. Previous researchers have actively explored combined application strategies for diverse evacuation signage types, investigating how to optimize the pattern design [32], static and dynamic status [33], and plane layout of evacuation signage [30]. These studies aim to maximize visibility and accessibility of evacuation guidance in complex emergency environments, ultimately providing comprehensive and effective guidance for evacuation. The combined use of dynamic and static signage offers significant potential advantages. In terms of flexibility in responding to dynamic scenarios, static signage is suitable for conveying stable information but struggles to adapt to sudden changes. Dynamic signage, on the other hand, can quickly respond to scenario changes through real-time updates of guidance and adjusting direction via rotating arrows. In terms of capturing passenger attention, static signage, due to its fixed information, is suitable for passengers to form stable memories. Dynamic signage, however, is more effective at capturing attention through visual dynamism and is suitable for conveying emergency or temporary information. Therefore, the combination of dynamic and static signs can adapt to the needs of different groups of people in different scenarios, not only to meet the dependence of daily commuting on fixed clues, but also to dynamically adjust the service of temporary passengers or special groups to achieve both efficiency and inclusiveness.

4.3. Recommendations for the Application of Virtual Reality Technology in Spatial Design

In traditional space research, the adjustment of spatial parameters not only requires substantial time, manpower, and material resources, but also faces challenges in achieving precise control due to susceptibility to interference from external environmental factors. In contrast, VR technology utilizes precise data acquired through field measurements and professional modeling software to construct highly accurate 3D models aligned with real subway stations, thereby creating immersive virtual scenarios with authentic spatial cognition effects. Given the complexity and uniqueness of subway station spatial environments, an increasing number of researchers are inclined to conduct evacuation signage-related studies within immersive virtual environments. During the experiment, researchers often use interaction devices such as eye-trackers and electroencephalographs (EEG) to accurately track participants’ behavioral paths and decision-making processes during evacuation [34,35,36]. This methodology makes it possible to quantify the specific impact of evacuation signage on the efficiency of evacuation under different conditions of visibility and accessibility. With this feature, VR technology provides innovative research ideas and methods for in-depth investigation of passenger wayfinding behavior and optimization of subway station spatial layouts. In light of this, researchers summarized the methodological flow of the application of virtual reality technology in spatial optimization, as shown in Figure 17.

Despite the results that have been achieved in this study, there are still some limitations. Although virtual reality technology can highly simulate the visual and spatial environments of subway stations, there are still gaps compared to real-world scenarios in terms of sight, sound, smell, touch, the actual sense of crowding, and the real pressure of sudden events such as smoke diffusion. These real-world environmental factors may interfere with passengers’ attention and decision-making processes, affecting the actual effectiveness of evacuation signage. Specifically, real-world stress factors can undermine evacuation effectiveness in multiple ways. High-decibel noise can distract passengers and slow down information processing, meaning that even if signage is optimally positioned, passengers may overlook it due to their excessive focus on escaping. Panic-induced herd behavior may cause passengers to follow the crowd rather than rely on signage for decision-making, further weakening the guiding role of signage. Smoke in scenarios such as fires can significantly reduce visibility. If smoke concentration reaches moderate levels, even signages hung at higher positions may become obscured by smoke and unrecognizable; conversely, signages at lower positions may maintain some visibility over short distances due to smoke’s upward movement, thereby reversing the effectiveness ranking of signages at different positions. Therefore, the conclusions drawn from this study primarily apply to low-pressure, normal-order wayfinding scenarios rather than high-pressure emergency evacuation scenarios. However, this also lays the foundation for future research considering pressure factors.

In the future, we may also consider incorporating the complex variable of crowd density, establishing experimental groups with different crowd densities to analyze the applicability of signage positions in dynamic crowd environments. Additionally, we could conduct partial experiments in subway stations during non-operational hours, combining virtual reality (VR) technology with augmented reality (AR) technology to superimpose simulated evacuation signage in real scenes. Using environmental simulators such as smoke generators and noise devices, we can create composite scenarios with visual obstructions and auditory interference to test the interference resistance of signage in adverse conditions. Alternatively, we could introduce physiological sensors in experiments to simulate and quantify the impact of stress levels on signage recognition efficiency, adjusting task difficulty to induce moderate stress, thus more comprehensively assessing the effectiveness of evacuation signage in a complex and real environment.

It is also worth considering the interaction between various design elements of evacuation signage, such as color, shape and symbols and location factors, while also taking into account the influence of cultural and linguistic contexts. Cultural differences can impact spatial cognition and reliance on signage. In most East Asian countries, where collectivist cultures dominate, passengers may be more accustomed to following unified signage and have a higher degree of reliance on it. The conclusions drawn from this study, based on a VR scene constructed at a subway station in Qingdao, China, may have some general applicability in similar cultural contexts. In Western countries with stronger individualistic cultural characteristics, passengers may be more inclined to explore the environment independently or rely on personal experience, resulting in lower reliance on signage. In such cases, the influence of signage position may weaken, and more personalized approaches may be required to achieve effective guidance. Additionally, different cultures have varying habits regarding the description of spatial directions. If the directional logic of signage conflicts with local cultural norms, its guidance efficiency may decrease even if the signage is well-positioned. To enhance the effectiveness of signage across different cultural and linguistic contexts, researchers recommend retaining optimal positioning while adjusting signage to align with local cultural characteristics. Exploring the optimal configuration of evacuation signage for different types of emergencies in diverse cultural contexts can provide more comprehensive theoretical support and practical guidance for future emergency management in subway stations.

5. Conclusions

This study utilizes VR technology to focus on the complex spatial structure and distinctive operational characteristics of subway stations, conducting a systematic analysis of the intrinsic relationship between the position of evacuation signage and passenger wayfinding efficiency. The findings contribute to the theoretical framework of evacuation guidance mechanisms. Experimental studies have confirmed that the optimal visual field position for evacuation signage is the hanging form. This shows significant advantages in real-world evacuation scenarios, effectively reducing evacuation time and substantially improving passenger wayfinding efficiency. These findings provide critical theoretical support for achieving rapid and orderly evacuation of personnel in subway stations. Secondly, signage positioned at wall columns contributes to maintaining relatively stable wayfinding efficiency for passengers during evacuation processes. However, during actual emergency evacuations, the high-density passenger crowds in subway stations can easily obstruct passengers’ visual field, which disrupts the continuity of evacuation information acquiring and thereby compromises evacuation efficiency to a certain extent. Finally, grounded signage, positioned within the lower visual field of human perception, requires passengers to deliberately look down for identification. This behavior not only diverts passengers’ attention during the evacuation process, preventing them from continuously monitoring dynamic environmental changes, but also significantly increases the risk of falls and injuries, thereby limiting overall guidance effectiveness. Nevertheless, within the multi-dimensional and complementary guidance system established in subway stations, wall-columned and grounded signage serve as critical auxiliary guidance measures. These elements retain indispensable supplementary functions in specific scenarios by providing alternative directional information, thereby ensuring passenger safety during wayfinding.

However, the building system of human society is rich and diversified, and civil buildings and public buildings constitute its main part. Beyond underground spaces exemplified by subway stations, people predominantly inhabit built environments characterized by functional complexity and morphological diversity, including office buildings, hospitals, educational institutions, and commercial complexes. There are significant differences compared to subway stations in terms of spatial layout, functional zoning, pedestrian flow volume, and movement patterns. However, whether it is a subway station, hospital, shopping mall, or other building, human wayfinding behavior follows the basic logic of first perceiving, then identifying, and finally making a decision. Based on this commonality in spatial cognition, the conclusions drawn in this study can be applied to other building types, but adjustments must be made in accordance with the specific spatial characteristics of each building. Differences in spatial complexity and the objectives of human activities across various buildings may alter the priority of signage placement. Subway stations, which primarily focus on one-way evacuation and transfers, have relatively linear paths, making hanging signage particularly effective for global guidance. In contrast, buildings with composite functions such as hospitals and shopping malls often feature multiple destinations and branching paths. In addition to global guidance, it is essential to reinforce signage at local nodes. Therefore, when applying the conclusions of this study to other building types, signage locations must meet requirements for visibility, coherence, and adaptability, while also flexibly adjusting height, information density, and combination methods based on the building’s spatial characteristics.

This study proposes an optimization strategy for evacuation signage positions in existing subway stations, which can significantly enhance evacuation efficiency. Furthermore, during the planning and design phase of future subway station construction, the position factors of evacuation signage should be fully integrated to enhance personnel safety during emergencies. Especially in the current prospect of building new intelligent subway stations, the conclusions of this study can provide a solid and reliable basis to help create a safer and more efficient subway evacuation system and effectively protect the lives and properties of passengers.