Evacuation Experiments Research Articles

Study of behaviors of passengers on ships and evacuation simulation method based on evacuation experiments

The accuracy and reliability of passenger evacuation simulation remain crucial in the research field of maritime pedestrian safety. However, a noticeable gap exists between current simulation results and actual pedestrian evacuation behaviors. This paper aims to enhance the precision of evacuation simulation result by gathering passenger behavioral data through on-ship evacuation experiments. The data collected during the evacuation are integrated with an evacuation simulation tool, and the pedestrian evacuation processes in both the experiment and simulation are analyzed and compared. The results reveal that the simulation tool has certain limitations, particularly in replicating passengers’ route selection and other specific behaviors. Nonetheless, when the evacuation experiment data is incorporated, there is a marked improvement in the congruency between the simulation and experiment results, compared to the use of the simulation tool alone. Therefore, incorporating drill data into the simulation scenario design presents a viable method to boost the accuracy of the simulation result. The findings of this study have practical implications for improving the accuracy of evacuation simulations and enhancing passenger safety on ships.

Individual route choice behavior in evacuation considering avoidance and phototropism: An experimental study

During an emergency evacuation in a building, participants' evacuation signs obedience behaviors have a significant impact on evacuation efficiency, while participants' possible phototropism and hazard avoidance behaviors occur during evacuation, which will affect evacuation efficiency. To investigate the effects of participants' phototropism and hazard avoidance behaviors on compliance with evacuation signs. One pre-experiment and four groups of evacuation experiments were conducted in the T-junction and Room of the building with a total of 170 participants. A fumes generator and searchlight to simulate light and fumes conditions under fire. Using the statistical analysis to explore the differences in the route choice behavior of evacuees under different conditions. The results show that: (1) In the natural state, participants showed significant phototropism and hazard avoidance. (2) Phototropism and hazard avoidance affect participants' obedience to evacuation signs. (3) In the room scene, participants demonstrated more obvious phototropism and hazard avoidance, even deviating from the evacuation sign's guidelines；(4) Participants showed some decision inertia during two consecutive decisions. Further combined with SHAP interpretable machine learning for analysis, it was found that (1) Scene conditions exist for phototropism and hazard avoidance in evacuees. (2) Fumes prevented participants' obedience to the evacuation sign, whereas light promoted participants' obedience to the sign. This indicates that evacuation signs need to be installed regarding the actual building structure. These findings can be used as theoretical references for building safety design, campus safety, and evacuation simulation tests.

Experiment and path optimization on Large-Space Multi-Area crowd evacuation

Abstract In large spaces where there are multiple evacuation zones, the flow patterns during evacuation are intricate, leading to congestion in emergency scenarios, and affecting evacuation efficiency. To address this, we proposed an optimization method for crowd evacuation paths in large-space multi-area environment settings under fire scenarios by utilizing multi-agent simulation. Using Pathfinder software, disorganized evacuation simulations are conducted within expansive multi-zonal settings to pinpoint areas with inadequate evacuation capabilities. Subsequent evacuation experiments are performed for verification, facilitating adjustments to agent-based model parameters for evacuees in fire scenarios, and carrying out organized evacuation simulations. Through meticulous simulation analysis, we aim to optimize the most efficient evacuation paths under organized responses in large-space multi-area settings. Results indicate that in emergency conditions due to fire, the evacuation speeds reach 1.8 m/s, and significant reductions in total evacuation time are achieved through simulation optimization, with the time required for the slowest individuals to evacuate decreasing from 240.32 s to 209.32 s. Furthermore, the flow rate per second at congested exits diminishes from 2.01 persons/s to 1.39 persons/s, highlighting the efficacy of the proposed path optimization method.

An experimental study on the impact of motivation and gender on pedestrian movement characteristics in a bottleneck flow

The influence of different motivations on pedestrian evacuation efficiency, like whether incentivizing faster moving or encouraging polite behaviors is beneficial to evacuation, and the potential existence of gender-based differences, still lacks clear answers. This study aims to narrow this gap by conducting a laboratory bottleneck evacuation experiment to investigate the influence of movement motivation and gender of pedestrians on evacuation efficiency. Our findings reveal that both bottleneck width and pedestrian motivation significantly impact the flow. For men, when the bottleneck width surpasses a threshold, high motivation increases the flow; below this threshold, it reduces the flow. For women, high motivation consistently leads to higher evacuation efficiency than normal motivation regardless of the bottleneck width. The time interval is significantly influenced by the bottleneck width, pedestrians’ gender and their interaction term. Increasing bottleneck width does not always lead to a decrease in density when men are highly motivated; while for women, wider bottlenecks alleviate congestion around the exit.

The Collection and Compilation of Small Group Data for Scenario Setting of Simulations and Experiments

The influence of small groups in evacuation cannot be ignored. However, the current simulations and experimental studies have oversimplified the settings of small groups and evacuation scenarios. A significant disparity exists between the findings of existing studies and real-world scenarios. This paper compiled data on the number and size of small groups and the location of small group members in built environments. Subsequently, a scenario with intricate functions and shapes was established, and finally, these data was employed in agent-based simulations. The data encompassed 50 small groups comprising a total of 111 members. These groups, ranging from 2 to 4 members each, exhibited spatial separations between members spanning from 1 m to 23 m. Simulation outcomes indicated a detrimental effect of small groups on overall evacuation. A significant and positive correlation was observed between the distance separating small group members and the escalation in evacuation time, total jam time, and evacuation distance. The data provides a foundation for configuring initial scenarios in small group evacuation experiments and simulations. The simulation results can provide a basis for hospital safety evacuation management.

Factors affecting evacuation in intellectually disabled children: An experimental study

Children with intellectual disabilities are considered a vulnerable group in evacuations during emergencies, yet their evacuation needs have not received sufficient attention. To bridge this gap, this study provides one of the very few experimental analyses of the evacuation of intellectually disabled children. Therefore, this study aimed to gather unique evacuation data for intellectually disabled children and identify factors influencing their evacuation process. Two evacuation experiments were conducted on 49 intellectually disabled children aged 7 to 16 from a special education school. Data collected through video recordings include recognition time, response time, pre-movement time, indoor travel speed, horizontal travel speed, and vertical travel speed. This study explored whether and how several key factors (e.g., drill announcement status, sex, age, floor level, and students' dependency behavior) affect the evacuation of intellectually disabled children. Results indicate that drill announcement status significantly affected the recognition time and pre-movement time in intellectually disabled children but did not impact their response time or travel speed. The floor level on which children were located significantly influenced their evacuation process, with children on higher floors showing longer response and pre-movement time and slower indoor travel speed and horizontal travel speed compared to those on the first floor. Additionally, the dependency behavior of these children on teachers significantly influenced their recognition time. Moreover, this dependency behavior tended to be prevalent among intellectually disabled children during emergency evacuations. These findings provide valuable insights for the design of educational buildings tailored to the needs of special populations, evacuation simulation modeling, and the development of building regulations.

Order–disorder phase transitions in front of the exit during human crowd evacuations

Human crowds often exhibit collective behaviors through self-organization, such as orderly queueing and disordered congestion at exits. Understanding the dynamic mechanisms behind these behaviors is crucial for pedestrian traffic management and the handling of mass crowds. Although current models and experiments have provided profound insights into ordered and disordered pedestrian groups separately, the transition mechanism from order to disorder within pedestrian crowds in front of exits remains unclear. In this study, a pedestrian queueing evacuation experiment was presented to demonstrate the phase transition process of pedestrian movement states as urgency levels in the environment change. We discovered the migration pattern of the phase transition point under evacuation control and identified different propagation modes of pedestrian velocity waves. Furthermore, a pedestrian motion model based on experimental data and observed phenomena was established. This model not only replicates the observed phenomena and regularities from the experiments but also reveals the quantitative phase transition mechanism of pedestrian movement under the influence of a single factor (i.e., urgency level or queueing ratio). Our findings hold significant implications for various domains, including pedestrian management, traffic control, and pedestrian dynamics.

An improved social force model for improving pedestrian avoidance by reducing search size

In order to solve the slow simulation speed of the social force model in complex indoor scenes, an improved social force model is proposed to reduce the iteration scale of obstacles. The proposed model takes into account the factor of a person’s field of view (FOV) in indoor scenes, and the acting force between pedestrians and obstacles in the social force model is improved to find the optimal number of obstacle iterations without affecting the pedestrian movement law in the traditional model, so as to accelerate the simulation speed. The self-built simulation platform and real evacuation experiments are used to compare the performance indicators of traditional and optimized models and to analyze the relationship between scene complexity and machine computing efficiency. The results show that the optimized model takes less time to predict the evacuation path than traditional models, and as the number of initial pedestrians and obstacles increases, the proportion of machine computation time between the two models will gradually increase. Second, the CPU computation time of the optimized model increases gently, indicating better stability. Therefore, reducing the obstacle iteration scale based on the pedestrian’s FOV factor provides an effective method to solve the slow simulation speed of traditional models.

A study on the arch mechanism of pedestrian evacuation and congestion alleviation strategies at building exits

The pedestrian flow capacity of building exits has always been a focus of research in evacuation field. This paper focuses on the congestion mechanism of pedestrian flow at exit bottlenecks through simulation and crowd experiments, and tries to propose strategies to alleviate evacuation congestion. Firstly, pedestrian flow evacuation experiments were conducted by setting up scenarios with exits in the middle and corners, and corresponding exit structure models were established using Massmotion software to analyze the congestion mechanisms of pedestrian flow at the exits through simulations coupled with the social force model. Secondly, the severity of congestion in the exit structure was quantified using the example of the middle exit. The width range of the pedestrian evacuation trajectory within 1 m before the exit was defined as the average chord length of the arch structure, and the area enclosed by the semicircular arch with this chord length was defined as the area with a relatively high instantaneous density of pedestrian flow. The results indicate that corner exits are more conducive to evacuation as they reduce the expected velocity of pedestrians near the center of congested areas. Both building width and the inflow rate of pedestrian flow, along with the number of pedestrians, demonstrate significant effects on the average chord length of the arch. The research results can be used to further supplement and improve the dynamic theory of pedestrian flow evacuation in buildings and provide a scientific reference for crowd congestion control at building exits.

Characteristics of pedestrian evacuation from narrow seated area considering exit failure: experimental and simulation results

Narrow seated spaces with multiple exits are prevalent structures in public buildings, underscoring the paramount importance of facilitating swift evacuation in such constrained environments. In this study, we first conducted evacuation experiments in a realistic narrow seated area. By manipulating different availability conditions for two exits located at the ends of the long aisle, we studied the effects of unpredictable exit failures, specifically, how the exit switch and explicit guidance influence the evacuation process. The movement characteristics are explored in several aspects, such as pre-evacuation time, aisle conflicts, speed, distance headway, evacuation time, and time lapse. We found that once people have encountered the exit failure condition, they would demonstrate enhanced preparedness in subsequent trials, leading to expedited exit switches. With explicit guidance, people would be less hesitant and move faster in the main aisle. The fine-grid cellular automata model is extended based on the experimental observations to verify its feasibility in simulating this process. Conflict resolution and turning behavior are incorporated for a more nuanced description. The effects of the exit switch are parameterized through the informed probability and the speed reduction.

Fast trajectory extraction and pedestrian dynamics analysis using deep neural network

Pedestrian evacuation dynamics research is crucial for safety engineering and crowd management. Accurate and fast extraction of pedestrian trajectories from experimental videos is essential for reliable evacuation data and effective strategy development. In this paper, we propose a novel method for extracting pedestrian trajectories from evacuation experiment videos based on deep learning techniques. The method consists of two modules: pedestrian detection and location prediction, which use Convolutional Neural Networks (CNNs) and Long Short-Term Memory (LSTM) networks, respectively. Experimental results show that our method achieves high accuracy with short time compared to traditional methods. Furthermore, our experiments demonstrate that our method can accurately extract information on pedestrian evacuation dynamics. The proposed method provides a fast and reliable approach to extracting pedestrian evacuation dynamics information, which has the potential to be utilized by researchers, urban planners, and emergency management personnel to develop more effective evacuation strategies, improve crowd management, and ultimately enhance public safety.

The study of self-organised behaviours and movement pattern of pedestrians during fire evacuations: virtual experiments and survey

In this study, Minecraft was adopted to investigate pedestrian behaviours and evacuation processes in non-fire (normal) and fire conditions involving multiple simultaneous players. We built a PaddlePaddle-based deep learning platform for the automatic recognition of pedestrian coordinates from the Minecraft background database. To verify the virtual experiments in Minecraft, we conducted a series of experiments under identical conditions in a real (physical) room and virtual environment on the Minecraft platform. The pedestrian behaviour, flow patterns and evacuation time in the Minecraft experiments fit well with the results of the real-life experiments, indicating the potential of Minecraft for realistically simulating evacuation processes. To further demonstrate the feasibility of using Minecraft for evacuation studies, we conducted a series of virtual evacuation experiments with different fire configurations and room geometries in Minecraft. The pedestrians were found to keep higher distance towards fire and greater detour angle during the fire emergency in a square room. Furthermore, in the scenario of a fire in a corridor, in addition to the aforementioned fire avoidance behaviours, the pedestrians exhibited orderly queuing and exited the room one by one, resulting in increased walking speeds and thus a faster evacuation for most pedestrians. Finally, one post-experiment survey was designed to assess participants' perceptions of the Minecraft environment and the pedestrian behaviours and related psychological factors. The results show that the Minecraft can generate considerate realistic real-life situations, and pedestrians chose self-organised behaviours including away from fire, along the wall and detour towards fire in both survey and evacuation experiments.

Experimental Study on Pedestrian Behaviors during Fire Emergency Conditions with Minecraft: Case Studies in a Classroom

The comprehension of the fire evacuation process is crucial for developing effective evacuation management strategies to enhance pedestrian safety. In this study, we construct a classroom with internal obstacles forming intersecting pathways in Minecraft, and conduct a series of virtual evacuation experiments involving multiple pedestrians to investigate the pedestrian behaviors. Case studies in a single-exit classroom demonstrated that normal obstacles and fire in the main evacuation path prompt pedestrians to detour, and pedestrians exhibit fire-avoidance behavior in advance during fire emergency. In the two-exit classroom experiments, normal obstacles have a limited effect on the exit choices of pedestrians, as they primarily choose the nearest exit. Pedestrians positioned in the center of classroom are influenced by their initial orientations, and some pedestrians opt for exits in their initial facing directions. The presence of fire has a greater influence on pedestrians’ exit choices, with most opting for exits away from the fire. Furthermore, during fire emergencies, some pedestrians engage in risk-taking behavior by choosing higher-risk paths in pursuit of a faster evacuation. These adventurous pedestrians proactively plan routes that maximize their distance from the fire and exhibit orderly queuing behavior. These findings are helpful to reveal pedestrian behaviors during fire emergencies.

Assessing potential tsunami vertical-evacuation practices: A study of four cases in Chile using virtual reality and GIS

Tsunami vertical-evacuation (TVE) can save numerous human lives when horizontal evacuation is not feasible. A key topic in TVE planning is the optimal placement of shelters, which is typically studied using a top-down approach. Recently, virtual reality (VR) has become an emerging tool to support disaster risk reduction and emergency response training. While VR has been applied to visualise tsunami flood zones and evacuation routes, no studies have used VR to assess the evacuees' shelter choices during emergencies or to examine the optimal 'siting' of TVE buildings. To address this gap, this study introduces a user-focused VR evacuation experiment to automatically collect the shelter preferences of 435 participants regarding existing buildings that might serve as TVE shelters in four case studies in Chile. Complementary, we used questionnaires to collect information about the participants' tsunami evacuation expertise, awareness, and the factors that led to their shelter selection. Lastly, we used GIS to examine the location characteristics of the selected buildings. The results showed that 50 buildings gathered at least two preferences, with a large diversity of heights (five to 34 storeys) and placements (from 52 to 2774 m from the coast). Height is the most critical factor when choosing a building for TVE (53.1 % of the preferences), and most of them are residential (62.7 %). Additionally, 67.2 % of the participants had experienced real-world evacuations, and the average evacuation rate during the last emergency in each of the case studies was 43.8 %. While most participants declared affirmative response rates about the perceived tsunami threat (above 60 %), low specific knowledge regarding evacuation plans might discourage protective actions in future tsunami emergencies.

Pedestrian flow through exit: Study focused on evacuation pattern

Experiments are conducted on the evacuation rate of pedestrians through exits with queued evacuation pattern and random evacuation pattern. The experimental results show that the flow rate of pedestrians is larger with the random evacuation pattern than with the queued evacuation pattern. Therefore, the exit width calculated based on the minimum evacuation clear width for every 100 persons, which is on the assumption that the pedestrians pass through the exit in one queue or several queues, is conservative. The number of people crossing the exit simultaneously is greater in the random evacuation experiments than in the queued evacuation experiments, and the time interval between the front row and rear row of people is shortened in large-exit conditions when pedestrians evacuate randomly. The difference between the flow rate with a queued evacuation pattern and the flow rate with a random evacuation pattern is related to the surplus width of the exit, which is greater than the total width of all accommodated people streams. Two dimensionless quantities are defined to explore this relationship. It is found that the difference in flow rate between the two evacuation patterns is stable at a low level when the surplus width of the exit is no more than 45% of the width of a single pedestrian stream. There is a great difference between the flow rate with the queued evacuation pattern and the flow rate with the random evacuation pattern in a scenario with a larger surplus width of the exit. Meanwhile, the pedestrians crowd extraordinarily at the exit in these conditions as well, since the number of pedestrians who want to evacuate through exit simultaneously greatly exceeds the accommodated level. Therefore, the surplus width of exit should be limited especially in the narrow exit condition, and the relationship between the two dimensionless quantities mentioned above could provide the basis to some extent.

Development and Implementation of a Smartphone Application for Tsunami Evacuation

Soft countermeasures help to reduce casualties in the event of tsunami, and the present study proposes the use of a smartphone application to aid with a smooth evacuation. The application, developed via Unity3D, which is a well-known game platform engine, aims to accurately reproduce the area of interest by showing the location of refuge places within the natural and constructed environment. A GPS-positioning system was incorporated in the application, such that users could easily identify their location and that of the designated refuge sites. This can be helpful for coastal areas that are popular sightseeing destination, especially for visitors who are unfamiliar with the region. The effectiveness of the smartphone application was further investigated by conducting evacuation experiments using human participants in Kamakura city, which is an area vulnerable to earthquake-induced tsunamis. Two groups were tested to compare the effectiveness of the developed smartphone application: One control group without the application, and the other with the application. The results revealed that the participants that were using the smartphone application took shorter evacuation times, travelled slower, were less likely to pay attention to tsunami evacuation signs, and covered less distance to arrive at the designated refuge places.

Human behaviour and evacuation time for large underground comprehensive buildings during fire risk process

Human factors play an indispensable role in process safety during fire risk process. Fire evacuation of underground comprehensive buildings is a critical consideration for practical importance. Here, evacuation experiments were conducted in the Guangzhou East Railway Station Plaza, a large underground comprehensive building in Guangdong, China, with the plaza occupants as the participants. The evacuation factors, namely evacuation time and evacuee-selected evacuation routes, were evaluated using empirical equations (the Togawa model and the Melinek and Booth model) and evacuation modelling software (EVACNET 4 and composite occupant evacuation model). The experimental data were used to verify the empirical equations and evacuation models and to modify specific empirical parameters. The effect on escape route judgement and selection under emergency situations was analysed. The results demonstrated that direct access to outdoor exits could considerably reduce evacuation time. The experimental N and ts were 1 person/m·s and 28 s, respectively. The total evacuation time was 281 s, and the EVACNET 4 and composite models’ simulations of safe evacuation were 285–290 s, which was consistent with the experimental data—validating the high accuracy and applicability of the models.

A comparative study on pedestrian flow through bottlenecks between flood evacuation and land evacuation

In a flood, people may be forced to evacuate in water until reaching a secure shelter. Understanding the characteristics of pedestrian flow in water is therefore of utmost importance. Some researchers studied individual walking speed in water, whereas it is still rare to study pedestrian and evacuation dynamics, e.g., characteristics of pedestrian flow through a bottleneck. In this paper, evacuation experiments through bottlenecks are conducted in 0.60-m-deep water, with 97 volunteers with an average age of 21.7 years participating. The results are compared with those on land. It is shown that the flow rate in water is lower than that on land, with a difference varying with bottleneck width b as ΔJ=0.60b−0.04ped/s. The average specific flow rate in water is 0.43ped/(m·s) lower than that on land. Based on the relation of time headway–width, density–width, and speed–width, threshold widths for the flow transition from “interrupted” to “flowing” stages are 1.3 m in water and 1.0 m on land respectively. Compared with that on land, during “interrupted” stage in water, stop behavior is more frequent, the speed near the exit is 0.22 m/s lower, and the speed-up point is closer to the exit. During “flowing” stage, continuity of the pedestrian flow at exit is inferior in water compared to that on land, a 0.2–0.3 m wider exit is required in water to achieve similar continuity on land. These findings may provide valuable insights into emergency management during flood evacuation.

Modeling exit choice behavior in airplane emergency evacuations

Understanding how people choose exits during the emergency evacuation of an airplane is fundamental to saving more lives after air accidents. Since exit choice is difficult to observe in an emergency directly, this paper utilizes a stated preference experiment to evaluate people's exit decisions in an airplane emergency evacuation. Responses were collected through online questionnaires conducted in China and the USA. Then, conditional and nested logistic models were applied, indicating significant effects from physical factors, social influences, and individual characteristics on exit choice in virtual airplane evacuations. First, a low-level emergency and traveling alone allow people to make broader exit choices. Then, people with professional aviation experience appear less likely to choose the backmost option than common air passengers, and a high BMI index positively impacts the selection of the backmost exit. Next, herding behavior was not dominant in the airplane emergency evacuation experiment, but people were found to avoid following others in rational decisions which were made in an unreal emergency. In addition, the strong impact of a family member's location on people's direction of movement was indicated in the results. Finally, the frontmost exit and movement towards the front of the aircraft were preferred in a crisis and when people travel with family. This study provides evidence of passenger exit choice behavior in airplane emergency evacuation and suggests that the seating position of family members should be taken into account to support more efficient and safer airplane evacuations.

Experimental and numerical evacuation study in tall office building

The duration of the evacuation process in case of emergencies in tall buildings is a crucial issue. Various factors have significant impact on the evacuation time, e.g. stairwell geometry, speed, density, pre-evacuation delay, number of occupants and their familiarity with the building, behavior, age, gender, physical abilities, relationships between people, etc. Although there are many studies on the subject, research results from Bulgaria and comparison with those from other places in the world are not yet available. Also, studies performed by researchers who have professional operational firefighting experience are very limited. In this study, two evacuation experiments in a new 107 m tall office building in Bulgaria were designed under the supervision of professional firefighting personnel. The first experiment was conducted to estimate the free movement characteristics under “normal” conditions for determining the average duration of the evacuation and the range of the vertical speed. Thirteen participants were instructed to move down from 17F to outside the building, following their normal habits. Their evacuation times differed in the range from 192 s–328 s, while the vertical speed is highest at the upper floors 0.35 m/s and reduce to 0.20 m/s just before the exits. The second experiment was an evacuation drill of 177 people from the floors they work in the building. The evacuation duration of this experiment for all these people from the building was in the range of 2–6 min. The study is also enriched through Pathfinder software simulations for verification of the results, where a good match was observed (6:26 min total evacuation time). Another simulation at the maximum occupants’ capacity of the building with 2195 people was also run (24:48 min for the total evacuation time). We believe that the benefits achieved on the differences in the real experimental data from different parts of the world can be valuable for validation of tall building evacuation models. In addition, the link between the evacuation drill findings and the professional firefighting considerations, problems and decision-making operations in case of fire is discussed.