Journal of Advanced Transportation

Journal of Advanced Transportation / 2020 / Article

Research Article | Open Access

Volume 2020 |Article ID 8791503 | https://doi.org/10.1155/2020/8791503

Hui Xu, Cheng Tian, Yang Li, "Emergency Evacuation Simulation and Optimization for a Complex Rail Transit Station: A Perspective of Promoting Transportation Safety", Journal of Advanced Transportation, vol. 2020, Article ID 8791503, 12 pages, 2020. https://doi.org/10.1155/2020/8791503

Emergency Evacuation Simulation and Optimization for a Complex Rail Transit Station: A Perspective of Promoting Transportation Safety

Academic Editor: Francesco Galante
Received11 Mar 2019
Revised26 Sep 2019
Accepted19 Oct 2019
Published07 Jan 2020

Abstract

Rail transit stations with multifloor structures have been built in many cities to intensively utilize land resources and facilitate lives of community. However, being overcrowded with passengers results in high risks during daily operation. In response, this study conducted an emergency evacuation simulation and optimization in the three-dimensional (3D) space of “complex rail transit stations” (CRTSs). The aim of the paper is to provide a methodology to determine effective emergency evacuation strategies for CRTSs. The Lianglukou Rail Transit Station in Chongqing, China, was used as a case study and the AnyLogic simulation platform employed for simulating emergency evacuations. An emergency evacuation theoretical framework was established. The emergency evacuation strategies, including evacuation routes and evacuation times, were determined based on the theoretical demonstration. Simulation and optimization of emergency evacuation in the Lianglukou station were conducted. Accordingly, four main simulation results were obtained: (1) Escalators/stairs and turnstiles are key facilities in the evacuation; (2) Effective guidance for the evacuation is necessary in the public space of the station; (3) Passenger aggregation nodes should be guided for balanced evacuation; (4) Removing metal barriers is a useful evacuation optimization measure. The proposed research method and framework can be used by other CRTSs in the establishment of emergency evacuation strategies and effective optimization strategies to promote safety of transportation system. The research findings are beneficial to passengers in helping them provide valuable emergency evacuation guidance.

1. Introduction

As a main component of modern transportation, urban rail transit conforms with the principle of sustainable development and has advantages of large capacity, high speed, energy saving, low pollution, convenient, and comfortable [1, 2]. Urban rail transit system could be subdivided into 7 categories, which are tram, light rail, rapid transit, monorail, commuter rail, funicular, and cable car, among which rapid transit includes underground, subway, tube, elevated, and metro (Mass Rapid Transit) [3]. To date, over 200 cities operate metros worldwide [4]. In China, metro and light rail are two major categories of urban rail transit, and over 5761 km of urban rail transit are in operation in 35 cities, ranking the first in the world [5]. The passenger flow volumes are constantly increasing and many rail transit stations have reached, or even exceeded, their designed long-term maximum passenger flow volume [6]. Rail transit stations have become highly populated spaces, especially transfer stations with multirail transit lines. In 2018, the urban rail transit passenger flow volume reached to 21.07 billion in China, with an increase of 2.59 billion (14%) compared to the previous year. The average daily passenger flow volume of Beijing and Shanghai rail transit are 10.54 million and 10.172 million, respectively. In October 2018, the daily passenger volume of Renminguangchang station in Shanghai reached to 0.76 million, which is the maximal value in the station passenger volume record [7].

With the expansion of the rail transit system, rail transit stations are subject to various kinds of potential risks due to passenger overcrowding [8]. Catastrophic accidents occasionally occur. On 10 February 2017, for instance, a Hong Kong metro fire injured 18 people; a Beijing metro elevator accident on 5 July 2011 killed 1 and injured 27 people; and a 19 February 2003 Daegu metro fire caused 198 deaths and 147 injured, to name just a few. The accidents not only influence the operation of one station, but also the whole rail transit transportation network, even the whole city. Emergency evacuation study is therefore a key issue to be considered to help guarantee the safety of rail transit stations.

In recent years, rail transit stations with multifloor structure and different functions integrated in 3D spaces have been built in a variety of Chinese cities, such as Beijing, Shanghai, Guangzhou, Chongqing, Chengdu [9, 10]. These stations consist of many spatial components, including various kinds of physical facilities, passengers, station staff, etc. [11] Complex relationships, that could be conceived as complex systems, also exist among different spatial components [12]. In this study, such stations are termed as complex rail transit stations (CRTSs). This study establishes a theoretical emergency evacuation framework of CRTSs by fully considering the structural complexity of the station. Based on the theoretical framework, the Chongqing Lianglukou rail transit station is used as a case study for modelling emergency evacuation with the AnyLogic simulation platform. The study is to identify a reasonable and efficient emergency evacuation route for pedestrian evacuation, which could provide reference for the safe operation of CRTSs.

2. Literature Review

Emergency evacuation was emphasized as an important task in building design [13, 14]. Evacuation routes, exits, stairways, passages, and turnstiles in rail transit stations are recognized as key facilities that directly influence evacuation route plans [15]. A crowd-aware environment design tool was proposed to analyze the optimal crowd flow-density relationships between crowd and environment [16]. The average minimum width of staircase of per person and the maximum upstairs speed were emphasized as two key parameters affecting the simulating evacuation [17]. A comprehensive study was made for the standard-based occupant evacuation design of Chinese metros, including evacuation to safe areas, evacuation time, and evacuation route [18]. A dynamic route planning problem in a restricted space evacuation was investigated through a heuristic algorithm [19]. Four crucial theoretical issues on the passengers’ evacuation are included, which are competitive vs. cooperative behaviours, proactive vs. reactive behaviours, route/exit choice, and symmetry breaking [20]. Pedestrian behaviours and their interactions in crowds were modelled, and particular interests are the self-organized patterns of motion and crowd exit-choice behaviours [21]. Taking into account the complex interactions between traffic and pedestrians, a systematic simulation-based multiattribute decision approach to route choice planning was developed and the uncertainties underlying pedestrian behaviours during an evacuation were modelled [22]. The main reasons and definitions of the crowd shockwave were proposed to study the correlation between people of different densities and stamping accidents [23]. The movement trajectory of pedestrian’s evacuation was focused under specific conditions (low visibility) [24]. The mechanism of buffer zone was studied to analyze the effects of different buffer zone lengths on evacuation time, density, pedestrian waiting time, and expected speed [25]. The movement characteristics of the mixed population and the wheelchair users at different widths of evacuation route were studied [26]. In the condition of shooting range fire, the effects of occupant response times and flame spread rate on emergency evacuation were studied [27].

Various emergency evacuation simulation approaches were proposed. A multiagent-based model was developed based on meta model for pedestrian simulation in metro stations and the model includes the three-level pedestrian agent model, the station environment abstraction model, and interactive rule base [28]. The evacuation time for a given number of evacuees was calculated with a multiagent-based simulation approach, while the evacuation bottlenecks in metro stations were also identified [29]. A theoretical framework for safety analysis of underground space was established based on previously reported fire accidents and empirical research into underground stations [30]. Besides, the waiting time during emergency evacuation in crowded transport terminals was studied through two evacuation models built by EXODUS and SIMULEX [31]. Different scenarios for simulating evacuations were considered based on insights from passenger behaviour surveys and a thorough psychological analysis of passengers [32]. Additionally, by combining 3D modelling with the agent modelling method, a multistandard ranking framework was established to quantitatively evaluate the evacuation factors affecting multistory buildings [33]. Combining evacuation exit modelling with agent-based simulations, the fire safety assessment of underground nuclear physics was conducted [34]. Path selection and evacuation strategies in case of complete or incomplete evacuation information were compared based on social force model [35]. An event-driven agent-based approach was proposed to model the fire emergency problem of ultra-high-rise elevator building [36]. The arching effect during emergency evacuation was studied for the university canteen, and the evacuation sequence of each channel is determined based on the principle of minimum evacuation time [37].

The structure characteristics of buildings have already been considered in the existing studies. In the ultra-high-rise building studies, the important role of refuge floor was emphasized. The elevator evacuation was a hot research topic for high rise buildings and was investigated by developing elevator evacuation models [36, 38]. A joint parameter modelling of building and crowds was established to study the density, path traces, speed in different activities and different crowd configurations [39]. The interactive system for computer-aided design optimization was contributed to improved building layouts [40]. Taking the mass rapid transit station as an example, the evacuation simulation process in the high rise building is divided into five stages and seven scenarios, which were analyzed with parameters, including walking speed, coefficient of flow rate, etc. [41]. The existing evacuation studies on the multifloor structure mainly concentrate on the high-rise buildings and most of the buildings are residential buildings or office buildings. These kinds of buildings show the features of multifloor and small public space on every floor. Only a few of research focus on the CRTSs. These kinds of buildings generally have less than 10 floors, large public spaces in every floor, and complex connections among different spatial components.

The above studies provide useful references for rail transit station emergency evacuation research in the aspects of physics structure deconstruction, pedestrian behaviours consideration, route choice, method choice, etc. Based on the existing studies on emergency evacuation, this research innovatively focuses on the emergency evacuation with considering the complex structure characteristics of the rail transit stations.

3. Materials and Methods

3.1. Characteristics of CRTSs

The mainly considered characteristics of CRTSs are multifloor structure and multitypical facilities in the space, which impact the inner space emergency evacuation directly and be the base of the emergency evacuation research. These two characteristics could be obtained and described clearly through field investigations. Thus, field investigations were conducted for some CRTSs, including the Xizhimen station in Beijing, the Hongqihegou station and the Lianglukou station in Chongqing, to clarify the structures and components of CRTSs.

3.1.1. Multifloor Structure of CRTSs

A multifloor structure can help satisfy the transfer demand from different rail transit lines. This generally includes platform floors and concourse floors. Beijing’s Xizhimen station, for example, is the transfer station for Lines 2, 4, and 13, and contains 6 floors comprising 3 platform floors and 3 concourse floors. The Chongqing Lianglukou station is constructed in 4 floors, of which B2 and B3 are concourse floors, while B4 and B5 are platform floors. The different floors are connected by 6 escalators/stairs.

3.1.2. Typical Facilities in CRTSs

(1) Escalators/stairs. Escalators/stairs provide vertical connections to different floors in 3D space. In emergency conditions, people on the platform floor should firstly go by the escalators/stairs to the concourse floor. All escalators from the platform to the concourse floor should be kept in operation, while the escalators in the opposite direction should stop operation and be used as fixed stairs, or reverse to the opposite direction [11]. The carrying capacity of the escalators/stairs is directly affected by their physical attributes, including width, length, gradient, and number, which should meet the requirements of an emergency evacuation [36].

(2) Turnstiles. Turnstiles are installed on the concourse floors and connect the CRTSs’ inner space to the outside environment. Generally, more than one turnstile is set in the concourse floor to meet the passengers’ different trip demands. The turnstiles are automatic ticket checkers in daily operation, but should be opened thoroughly for evacuation in an emergency. The capacity of the turnstiles is determined by their number and width, which should be designed based on the prediction of the passenger flow volume in daily operations and emergency evacuation. The turnstiles should also be clearly labelled in all directions to provide guidance to passengers.

(3) Screen doors. Screen doors in every platform floor are used to separate the platform and train operation area. The screen door area is always a concentration point of passenger flow for it directly connects to the passengers’ waiting area. Screen doors should be completely opened for evacuation in the event of an emergency.

(4) Metal barriers. Metal barriers are used on both the platform floors and concourse floors to split large volumes of passengers and guide the directions of passenger flows. They are placed in the screen door area or in front of the escalators/stairs to avoid crowding. Their adjustment should be considered in the event of an emergency to prevent becoming a bottleneck during an evacuation.

3.2. Case Study
3.2.1. Task

The selected study site should have representativeness for the CRTSs. The existing emergency simulation studies emphasis on crowded passengers, while this paper mainly focuses on the complex structure characteristic of the rail transit station. Accordingly, the study site should meet 3 criteria, that (1) daily crowded passengers, (2) complex multifloor structure (more than 2 floors), and (3) occupying with more than one rail transit lines.

3.2.2. Study Site

The Lianglukou rail transit station in Chongqing, China conforms to the above 3 criteria. By December 2018, a total of 10 lines are in operation in Chongqing, incorporating 178 stations and 13 transfer stations [42]. Of the 13 transfer stations, the multifloor complex Lianglukou rail transit station is the transfer station for Line 1 (metro line) and Line 3 (light rail line), which occupy some distinctive stations, including 5 commercial centers, 3 railway stations, and the Chongqing airport. This, together with Chongqing’s 2018 urban population of 20.32 million, has led to an extremely large volume of transfer passengers, with a daily average of nearly 170,000 people in the 4 floors of the Lianglukou station [43], making the provisions for emergency evacuation of vital importance.

3.2.3. Variables of the Site

Five field investigations were conducted of the station to clarify its complex structure and obtain the passenger aggregation nodes for every floor. The station contains 4 floors: B2, B3, B4, and B5 floor. The investigation found high-density passenger volumes to be mainly concentrated on B3, B4, and B5. B3 is the concourse floor and includes 4 turnstiles, which connect to exits 5, 6, and 7, and unified exit 1–4 (the independent exits 1, 2A, 2B, 3, and 4 are on B2 floor). B4 is the platform floor of Line 1, with 24 screen doors on each side of the platform and hence a total of 48 screen doors. B5 is the platform floor of Line 3, with 16 screen doors on each side and hence a total of 32 screen doors. B3, B4, and B5 are interconnected with a total of 5 escalators/stairs. Some metal barriers are used on each floor to guide the directions of passenger flows. The dimensions of the three floors were also measured as part of the fieldwork (Table 1).


FloorLength (m)Width (m)Area (m2)

B3103.815.01557
B496.012.01152
B5107.412.01289

3.3. AnyLogic Simulation Platform

The emergency evacuation simulation for the 3D space of the station was carried out using the AnyLogic simulation platform, which was developed in St. Petersburg, Russia, a forward-looking simulation software based on the Java programming language and social force model, which can simulate the evacuation process of the crowd during emergency evacuation. The platform shows the characteristics of visibility, interactivity, and capacity of excellent analysis and optimization, which provides a rich animation of model execution and handles randomness. A virtual prototyping environment for large and complex systems with continuous, discrete, and hybrid behaviour is also provided [44]. The platform is suitable for complex system modelling within a random environment and to identify optimal operational strategies of system control based on risk analysis [45].

The platform consists of autonomous adaptive objects called “agents.” An agent exercises its autonomy through actions in the situated environment. The 3 essential characteristics of “agents” in the agent-based modelling are: (1) An agent is a self-contained, modular, and uniquely identifiable individual; (2) An agent is autonomous and self-directed. Agents’ interactions generate information, which influence the agents’ actions. For example, to avoid collisions with other agents, agents may choose accessible routes rather than the shortest route, and also, may chose accessible exits rather than the closest exit in the emergency evacuation; (3) The state of agents vary over time and the behaviours of agents are conditioned on the state [46].

The simulation follows four steps:(1) Establishment of the physical space. The facility module and waiting area module are set to simulate the physical space. The facility module includes trains, escalators/stairs, screen doors, turnstiles, security inspection equipment, ticket counters, etc. The waiting area module is the area used by passengers waiting for trains, and includes such facilities as columns, metal barriers, passages, etc.(2) Definition of passenger agent behaviours. The passenger agent behaviours include taking escalators, using stairs, waiting for trains, distributing in different areas, and choosing evacuation routes.(3) Programme editing. Based on the Java programming language, the simulation programme was edited. Such as formula “inject( )” is invoked by the command “passSouceOutBoard.get(i).inject( )” to set passengers in physical space. For example, 50 passenger agents are set through every screen door by the targetline setting “collectionDoor.get(index),” which means the set number of passenger agents going through every screen door is 50. The index denotes every specific door, such as targetlinedoor 1, targetlinedoor 2, … etc.(4) Operation and analysis of the simulation.

3.4. Social Force Model

The social force model was used as the crowd steering technique in this paper and be the base of AnyLogic simulation platform. The walking speed of evacuees in this model is driven by the motive force, the interaction among the evacuee, and the inter-action forces between the evacuee and obstacles.

The social force model is shown as follows,

In this model, is the quality of agent ; is the desired velocity of agent ; is the direction of desired speed of agent ; is the actual velocity of the agent ; is the relaxation time of agent ; and is the interaction forces from others exerted on agent ; is the interaction forces from walls exerted on agent .

is the interaction force between the agent and agent :

This function describes the psychological tendency of two agents by a repulsive interaction, where and are constants. Pedestrian and pedestrian keep a certain distance. denotes the distance between the pedestrians’ centers, and is a normalized vector pointing from pedestrian to . The pedestrians touch each other if their distance is smaller than the sum of their radius. In this case, the model assumes two additional forces inspired by granular interactions, which are essential for understanding the effects in dense crowds: counteracting the body compression and impeding the relative tangential motion if pedestrian comes close to .

is the interaction force between the pedestrian and the wall , and the formula is similar as :

is a small random force. Its size follows the normal distribution and direction is perpendicular to the direction of velocity [25].

4. Theoretical CRTS Emergency Evacuation Framework

The theoretical emergency evacuation framework consists of emergency accidents, accident spots and evacuation routes, evacuation time, and walking speed [47, 48]. The four aspects are necessary and essential to the emergency evacuation planning and implementation. The emergency accidents are the background or scenarios of the evacuation. The accident spots and evacuation routes are key elements for the evacuation planning and practice guidance. The evacuation time calculation method is necessary for the assessment/validation of the evacuation simulation. The walking speeding is a significant variable for the evacuation time calculation and the whole evacuation simulation process. Combined with the general characteristics of CRTSs, the 4 aspects of the framework were applied to the CRTSs emergency evacuation. The framework is shown in Figure 1.

4.1. Emergency Accidents

Past episodes indicate that public safety accidents provide a heavy threat to CRTSs. The main types of public safety accidents are fire accidents, terrorist attacks, and poison-gas accidents [49]. The common features are sudden occurrence, serious influence, and a small probability [50]. These public safety accidents are mainly considered in this research.

4.2. Accident Spots and Evacuation Routes

Emergency accidents can happen in different spots of the station. These involve three main situations: (1) emergency accidents in a station train (a train in emergency accident stopped at the station), (2) emergency accidents on a platform floor, and (3) emergency accidents on a concourse floor. When an emergency accident happens, the station enters an emergency state, and passengers need to be evacuated immediately. The emergency routes of the first situation are with the longest distance. In this situation, passengers firstly escape from the train and then go through the screen doors, platform floor, escalators/stairs, concourse floor, turnstiles, and finally to a safe spot outside the station. In the second situation, when accidents happen in the platform floor area, passengers should use the escalators/stairs, arrive at the concourse floor, and then go to the turnstiles to reach the station’s exits. In the third situation, when accidents happen in the concourse floor area, people go directly to the turnstiles. If the emergency accident happens near a turnstile, the emergency evacuation route has two conditions. One is that people directly escape from this turnstile providing the turnstile is opening and available for escape. The other is that people run to other turnstiles to exit the station.

4.3. Evacuation Time

Based on observation and analysis of the passengers’ behavioural features, the evacuation time is considered to comprise 5 parts, and they are listed as follows.

4.3.1. Response Time

The response time is preaction time, which is the time from the occurrence of an accident to starting an escape action. According to the illustration in literature [11], the response time is considered 1 min.

4.3.2. Time for Reaching the Escalators/stairs

where denotes the distance from the farthest point to the escalators/stairs and is the passenger walking speed. It is easy for people to find out escalators/stairs when going out of the screen doors. In emergency evacuation, people may go by the farthest escalator/stair rather than the nearest one due to crowding.

4.3.3. Time for Using the Escalators/stairs

where denotes the number of train passengers and is the number of people on platform, including passengers and staff, and is the passing capacity of the staircase/escalator.

4.3.4. Time for Arriving at Safe Spots

where is the length of the passage from the escalators/stairs to the safe spot.

4.3.5. Nonuniform Passage Deviation Time

People may gather in one passage during evacuation. Due to the distance between passages, an unbalanced gathering in the passages may delay the progress of evacuation. The maximum time deviation caused by an unbalanced gathering can be determined by

where is the distance between two different passages [11].

The total evacuation time is therefore given by

4.4. Walking Speed

Pedestrian walking speed has been explored in many studies. Normally free walking speed can be regarded between 1.2 m/s and 1.8 m/s [51]; an experiment by the National Traffic Safety and Environment Laboratory (Tokyo) of 28 persons walking on a crosswalk obtained an average walking speed of 1.2 m/s; while Imanaka and Yasui (2003) investigated 1444 pedestrian walking speeds on a crosswalk in front of a railway station during commuting hours and reported the average walking speed to be 1.7 m/s [52]. The comfortable walking speed ranged from 1.27 m/s (for women in their seventies) to 1.46 m/s (for men in their forties), while the maximum walking speed range from 1.74 m/s (for women in their seventies) and 2.53 m/s (for men in their forties) [53]. The walking speeds of pedestrians at the time of accidents are higher than on a crosswalk. Matsui et al. (2013) obtained the average pedestrian walking speed of 2.0 m/s through the investigation for 101 near-miss accidents [54]. According to the above studies and experiments, the passengers walking speed is determined as 2.0 m/s in this study.

5. Results

5.1. Coding of Facility Nodes

The detailed field investigation indicated screen doors, escalators/stairs, metal barriers, and columns are passenger aggregation nodes and coded by their first capital letter. For example, the screen doors in the left side of the B5 floor were coded as B5-LD1, B5-LD2…, and the escalators/stairs, metal barriers, and columns in B5 floor were coded as B5-ExxSxx, B5-Fxx, and B5-Cxx, respectively. Moreover, each floor in the CRTSs was classified into three areas according to the length data. The areas were coded as BX AX. The layout plan for B5, for example, is shown in Figure 2.

5.2. Emergency Evacuation Routes in Different Risk Scenarios

The simulation of the emergency evacuation was conducted for an accident occurring in a stationary train on B5 floor, as evacuation routes in this scenario are the longest. The different locations of accidents, including areas A1, A2, and A3, lead to different evacuation routes. The evacuation routes may go through screen doors, passages, metal barriers, escalators/stairs, and turnstiles, and end at the station exits. Taking the escalators/stairs as the key points, and suppose an emergency accident happens in a stationary train on B5 floor, the evacuation routes from B5 floor are shown in Table 2, and the evacuation times being calculated according to Equation (9) in Section 4.3.


No.Accident occurring areaEvacuation routesEvacuation time

1B5 A1B5 E3S3 → B4 E1S1 → B4 E2S2 → B3 E2S2 → Exit7270.6 s
2B5 A1B5 E3S3 → B4 E1S1 → B4 E3S3 → B3 E1S1 → Exit5276.3 s
3B5 A1B5 E2S2 → B3 E3S3 → Exit5256.2 s
4B5 A1B5 E1S1 → B3 E4S4 → Exit7243.2 s
5B5 A1B5 E1S1 → B3 E4S4 → B3 E5S5 → B2 E1S1 → Exit1-4293.5 s
6B5 A2B5 E3S3 → B4 E1S1 → B4 E2S2 → B3 E2S2 → Exit7288.5 s
7B5 A2B5 E3S3 → B4 E1S1 → B4 E3S3 → B3 E1S1 → Exit5299.3 s
8B5 A2B5 E2S2 → B3 E3S3 → Exit5263.3 s
9B5 A2B5 E1S1 → B3 E4S4 → Exit7232.2 s
10B5 A2B5 E1S1 → B3 E4S4 → B2 E1S1 → Exit1-4270.6 s
11B5 A3B5 E1S1 → B3 E4S4 → Exit7218.0 s
12B5 A3B5 E1S1 → B3 E4S4 → B2 E1S1 — Exit1-4256.3 s

The bold value (299.3 s) is the calculated maximum evacuation time.
5.3. Emergency Evacuation Simulation

Through field investigation, the number of passenger agents reaches to the maximal level when train arrives. In this condition, the average passenger number of every screen door is around 40. In daily operation, the two types of train are 6-marshalling and 8-marshalling, and only one side of the screen doors are open. Thus, generally, 640 = 240 or 840 = 320 passenger agents will enter the B5 floor when the train arrives. In the simulation, the maximal utilization of all the screen doors in both sides is considered, with the initial number of passenger agents on the B5 floor is determined as 2320 = 640.

In the simulation, the evacuation routes are the 12 routes in Table 2, which have been input in the AnyLogic simulation platform and the social force model has been used. The number of people going through every route is not determined, which is related with the accessibility of different route in the simulation.

Table 2 shows that the calculated maximum evacuation time is 299.3 s through the 7th evacuation route, while the minimum evacuation time is 218.0 s through the 11th evacuation route. The total simulation process lasts 308 s, which is longer than that along the routes listed in Table 2, that because the simulation time is the comprehensive evacuation time for people evacuating through the whole 12 routes from B5 floor, and includes not only the moving time, but also the evacuees’ waiting time in front of escalators, stairs, and turnstiles. The emergency evacuation time (308 s) could meet the “6 min” requirement of the Chinese Code for Safe Evacuation Design of Metro Station. It also indicates that the proposed evacuation routes are acceptable.

The relationships between the number of people on the B5 floor and evacuation time is presented in Figure 3. It shows that at the 20th second, people all evacuate from the train to the B5 floor. And at about 170th second, people all evacuated out of the B5 floor.

5.3.1. Identification of Distinctive Time Points

The distinctive time points during evacuation simulation were determined based on two serious problems, which are serious congestion and evacuation direction confusion. For serious congestion, some facilities, due to the width and number, would gather large numbers of people and shape as bottleneck in evacuation, such as escalators/stairs and turnstiles. Accordingly, the 28th second and 58th second are determined as distinctive time points. In addition, from the platform floor to the concourse floor, people go from a constrained space to a wide space and would be confused with suitable directions to safe areas. Therefore, the 170th second is determined as the third distinctive time points. The simulation snapshots are shown in Figure 4.

(1) Simulation at the 28th second. When an emergency accident happens on the B5 floor, the passenger agents go through escalator/stair to escape to the safety areas. At the 28th second, the 640 passenger agents gather at the entrances to the escalators/stairs in B5 floor—B5 E1S1, B5 E2S2, and B5 E3S3—which encounter serious congestion in the shape of an arch, especially around B5 E1S1 and B5 E2S2, as highlighted in yellow circles in Figure 4. According to statistics, totally 200, 168, and 33 people around the B5 E1S1, B5 E2S2, and B5 E3S3 at the 28th second, and the average number of people in per square meter of these three points is 2.581, 2.224, and 0.85. The crowd of people may further cause a stampede or other risk events. The escalators/stairs are therefore key facilities in the evacuation, as the evacuation capacity is directly related to the length and width.

(2) Simulation at the 58th second. At the 58th second, the first person (in yellow circle) is evacuated to the safe area through Exit 7 on B3 floor, as shown in the simulation scenario in Figure 4. The station is in a state of emergency and all the turnstiles are opened. From the B5 floor to the safety areas on B3 floor, the passenger agents go through screen doors, escalators/stairs, passages, and turnstiles. There is a large space on B3 floor and the evacuated people may start to push and stampede during evacuation. Thus, the number and width of turnstiles are major elements deciding the evacuation time.

(3) Simulation at the 170th second. As shown in Figure 4, all passenger agents have been evacuated to the concourse floors, B2 and B3, at the 170th second. Conformity psychology is a common phenomenon during the process of emergency evacuation [55], as most passenger agents tend to be reactive, cooperative, and follow the crowd with the same direction-attraction power. This may be because people do not clearly know the location or direction of the safety area. Thus, effective guidance for the evacuation, such as signs indicating the direction of the safety areas in the concourse floor, is necessary in the public space of the station.

(4) Simulation comparison to other studies. In this research, the 640 passengers in the Lianglukou Station are evacuated in 308 s. The average evacuation time is 0.48 s/person. Comparatively, (1) taking the two-floor canteen as an example, 268 people were simulated for evacuation, and the total evacuation time was 179.3 s (0.67 s/person) [56], (2) in the two-floor under-ground island station, considering evacuation in different escalator operation strategies and the movement state of persons on escalator, 554 pedestrians were evacuated between 195 s and 242 s (0.35 s/person to 0.43 s/person) [11], (3) taking a multifloor transfer metro station as the background, 2660 people were simulated for evacuation, and the total evacuation time was 750 s (0.28 s/person) [57]. Thus, compared to other simulation research, the simulation evacuation time of the Lianglukou station (0.48 s/person) is in the scope of the listed existing studies (0.28 s/person to 0.67 s/person). Thus, the evacuation simulation in this research is reasonable in this degree.

5.3.2. Analysis of Passenger Flow through Escalators/stairs

The escalators/stairs are the key passages for evacuation in the 3D space of multifloor structures, and the analysis of passenger flows through these points can provide information concerning the passage rate and passenger flow aggregation points, which is useful for offering guidance in the evacuation. There are three escalators on B5 floor, coded B5 E1, B5 E2, and B5 E3, each with an accompanying staircase correspondingly coded as B5 S1, B5 S2, and B5 S3, as shown in Figure 2. Figure 5 shows the number of agents passing through the different escalators/stairs.

As mentioned earlier, all the agents have evacuated to the concourse floor (B3) by the 170th second. According to Figure 5, the total number of agents having passed through the escalators/stairs at the 200th second is 223 + 138 + 77 + 92 + 81 + 29 = 640. The agents evacuating from B5 E1, B5 S1, B5 E2, and B5 S2 reach the B3 floor directly, while agents evacuating from B5 E3 and B5 S3 need go further, through B4 E2 and B4 E3 to the concourse floor. B5 E1 links B5 and B3 floors directly and thus has a higher evacuation ability. Most agents choose B5 S1 to evacuate. The number of agents passing through B5 E2 ranks the second. Agents going through B5 E2 from B5 to B3 floors need an immediate transfer at the B4 floor, which causes crowded clusters and extends the evacuation time. Agents who evacuate through B5 E3 should firstly arrive at B4 floor, then to B4 E2 or B4 E3, before arriving at B3 floor. The transfer from B5 E3 to the B4 escalators takes more time. Thus, only a few agents pass through B5 E3 and B5 S3.

It can be concluded, therefore, that (1) B5 E1 and B5 E2 are key facilities for emergency evacuation, and (2) the evacuation numbers through the escalators are unbalanced due to accessibility difficulties of the evacuation routes. Based on these results, further guidance for evacuation can be formulated.

5.4. Evacuation Simulation for Nonhomogenous Agents

Accounting for different kinds of crowds, further experiments on the AnyLogic simulation platform have been conducted, that considering 20% the aged/disabled people integrate into the evacuees, respectively. The speeds of these two kinds of people are determined as 1.58 m/s and 1.41 m/s [26], respectively. The simulation results of the evacuation time are shown as Table 3. The results present that the evacuation time is longer than the simulation for the homogenous agents (308 s). These conditions should be considered in practice management.


No.Experimental agentsEvacuation time (s)

1100% original passengers308 s
220% the aged + 80% original passengers341 s
320% disabled people + 80% original passengers373 s

5.5. Optimization Measures for the Emergency Evacuation

Due to the possibility for the metal barriers to create a bottleneck during an emergency evacuation [58], the optimization strategies are mainly concerned with removing metal barriers in different places. Six experiments were conducted to compare the optimization effects. According the simulation, when an accident happens in a train on B5, the evacuation time is 308 s without optimization. The first experiment was to remove all the metal barriers in B5. The second and third experiments were to remove parts of metal barriers in B4. B4 E1S1 was included in the two experiments as it is a necessary way for people to enter B4. The fourth, fifth, and sixth experiments are to remove metal barriers in different areas of B3. The results are shown in Table 4.


No.ExperimentEvacuation time (s)Reduced time (s)

1No optimization3080
2Remove all metal barriers on B5 floor28028
3Remove all metal barriers on B4 E3S3 and B4 E1S129117
4Remove all metal barriers on B4 E2S2 and B4 E1S129612
5Remove all metal barriers on B3 A129414
6Remove all metal barriers on B3 A226048
7Remove all metal barriers on B3 A329018

This indicates that removing all the metal barriers in B3 A2 could reduce the evacuation time to the maximum degree of 48 s. The metal barriers of B3 A2 are mainly concentrated in front of B3 E2S2, which is directly linked to Exit 5. Moreover, removing all the metal barriers in B5 floor is also a useful optimization strategy, as it could reduce the evacuation time by 28 s.

Removing metal barriers in the occurrence of an emergency condition requires lots of efforts of staff members. During the removing process, collisions between metal barriers and people may happen and lead to further obstacles. Thus, the implementation of removing metal barriers should consider the evolution results. Removing metal barrier exercises should be implemented regularly for effective and efficiency actions. Staff should command the removing methods.

6. Conclusion

As a main component of modern transportation, rail transit is highly popular in many Chinese cities. Rail transit stations with a multifloor structure integrated in 3D space have become greatly populated spaces. The passenger flow volumes of many stations are reaching or even exceeding their designed long-term maximum capacity. As rail transit stations are urban public spaces, their overcrowding makes them subject to threats from various emergency accidents. Research involving emergency evacuation simulation and optimization based on their complex structure is therefore significant for the safe operation of the stations.

The main contents of this study comprise the illustration of the complex structural characteristics of CRTSs, the establishment of an emergency evacuation theoretical framework of CRTSs, an emergency evacuation simulation for the case station, and the analysis of potential emergency evacuation optimization strategies. Using Chongqing’s Lianglukou rail transit station as the case study, the proposed evacuation strategies could meet the “6 min” requirement of the Chinese Code for Safe Evacuation Design of Metro Station, which provide a useful reference for the safety management of the Lianglukou station. From this case, four main simulation results were obtained:(1) Escalators/stairs and turnstiles are key facilities in the evacuation. The evacuation capacity of the station is directly related to the length, width, and number of escalators/stairs, especially at the beginning stage of the evacuation. Besides, the number and width of turnstiles are also a main element deciding the evacuation time. These could be illustrated by the evacuation situation in the Lianglukou station at the 28th and the 58th seconds.(2) Effective guidance for the evacuation is necessary in the public space of the station. The importance of effective guidance could be resulted from the evacuation phenomenon in the Lianglukou station at the 170th second. Effective guidance in the emergency evacuation includes signs indicating the direction of the safety areas, voice broadcast, evacuation commands from professional staffs or policemen, etc., which change dynamically according to the situation and aims to tell people the location or direction of the safety area [59]. Effective crowd guidance can improve egress efficiency, occupant survivability, and mitigate or prevent undesirable consequences such as stampeding or blocking [60].(3) Passenger aggregation nodes should be guided for balanced evacuation. Passenger aggregation nodes should be found through simulation experiment. In the Lianglukou station, the B5 E1 and B5 E2 are key facilities and passenger aggregation nodes for the emergency evacuation of the station, and guidance should be developed to balance the number of evacuees using each escalator/stair.(4) Removing metal barriers is a useful measure to optimize evacuation time. Removing metal barriers would be a useful strategy to shorten evacuation time, but the implementation should be exercised regularly in daily operation.

For the proposed emergency evacuation research methods, the AnyLogic simulation platform and the social force model could also be used in emergency evacuation for some other CRTSs, such as railway station, commercial complex, airport terminal, etc. In addition, for the established emergency evacuation framework, the four aspects of the framework could be adopted through combining with the characteristics of the research CRTSs. Firstly, the emergency accidents could be determined according to the statistics of accidents of the research CRTSs. Secondly, the accident spots and evacuation routes should be analyzed based on the structure of the research CRTSs and some scenarios should be considered. Thirdly, the evacuation time calculation method could also be used in evacuation time calculation for some other CRTSs. Fourthly, the review for the walking speed studies and the consideration for the walking speed could also be adopted in other studies. Thus, though the established framework could not be used in the emergency evacuation research for other CRTSs directly, the framework provides meaningful and useful references. The proposed simulation method is helpful for other CRTSs to determine reasonable emergency evacuation strategies according to the evacuation process and the structure/passenger characteristics of the case, which promote the safe operation of the urban rail transport system. Besides, the research could also supply valuable emergency evacuation guidance for passengers.

In the future research, three aspects will be mainly focused to further improve the simulation, including (1) collecting the real-world data and compare to the model output, (2) adding the subjective factors, such as psychological factors and cognitive factors, to the evacuation methodology, and (3) considering more emergency simulation scenarios, such as bottleneck in safety exits or passageways.

Data Availability

The data used to support the findings of this study are available from the field investigation upon request.

Conflicts of Interest

The authors declare that there are no conflicts of interest regarding the publication of this paper.

Acknowledgments

This work was supported by National Natural Science Foundation of China [grant number 71801026]; the MOE (Ministry of Education in China) Project of Humanities and Social Sciences [grant number 17YJC630189]; the Chongqing Research Program of Basic Research and Frontier Technology [grant number cstc2017jcyjAX0359]; and the China Scholarship Council Fund (grant number 201807845019).

References

  1. H. Xu, L. Jiao, S. Chen, M. Deng, and N. Shen, “An innovative approach to determining high-risk nodes in a complex urban rail yransit station: a perspective of promoting urban sustainability,” Sustainability, vol. 10, no. 7, p. 2456, 2018. View at: Publisher Site | Google Scholar
  2. F. Qin, X. Zhang, and Q. Zhou, “Evaluating the impact of organizational patterns on the efficiency of urban rail transit systems in China,” Journal of Transport Geography, vol. 40, pp. 89–99, 2014. View at: Publisher Site | Google Scholar
  3. Wikipedia, “Urban rail transit,” https://en.wikipedia.org/wiki/Urban_rail_transit#Rapid_transit, January 2, 2019. View at: Google Scholar
  4. METROBITS.ORG., “World metro database,” http://mic-ro.com/metro/selection.html, February 23, 2018. View at: Google Scholar
  5. General Office of the State Council in China, “Opinions of the general office of the state council on further strengthening the planning and construction management of urban rail transit. website of the central people’s government of the people’s Republic of China,” http://www.gov.cn/zhengce/content/2018-07/13/content_5306202.htm, September 16, 2019. View at: Google Scholar
  6. C. Shi, M. Zhong, X. Tu, T. Fu, and L. He, Experimental and Numerical Analysis on Deep Buried Metro Station Fires, Science Press, Beijing.
  7. China Urban Rail Transit Association, “2018 China’s Urban Rail Transit Annual Statistical Analysis Report,” http://www.camet.org.cn/index.php?m=content&c=index&a=show&catid=18&id=16219, September 20, 2019. View at: Google Scholar
  8. W. Chen, Y. Zhang, M. T. Khasawneh, and Z. Geng, “Risk analysis on Beijing metro operation initiated by human factors,” Journal of Transportation Safety & Security, pp. 1–17, 2018. View at: Publisher Site | Google Scholar
  9. The State Council of China, “Development plan of modern integrated transportation system in 13th Five-Year,” http://www.gov.cn/zhengce/content/2017-02/28/content_5171345.htm, January 30, 2018. View at: Google Scholar
  10. Chinahighway.com, “Integrated transport hubs are constructed in 7 cities,” http://www.chinahighway.com/news/2017/1087394.php, January 30, 2018. View at: Google Scholar
  11. C. Shi, M. Zhong, X. Nong, L. He, J. Shi, and G. Feng, “Modeling and safety strategy of passenger evacuation in a metro station in China,” Safety Science, vol. 50, no. 5, pp. 1319–1332, 2012. View at: Publisher Site | Google Scholar
  12. Y. Zheng, B. Jia, X. G. Li, and R. Jiang, “Evacuation dynamics considering pedestrians’ movement behavior change with fire and smoke spreading,” Safety Science, vol. 92, pp. 180–189, 2017. View at: Publisher Site | Google Scholar
  13. M. Usman, D. Schaumann, B. Haworth, and G. Berseth, “Interactive spatial analytics for human-aware building design,” in 11th Annual International Conference on Motion, Interaction, and Games, ACM, New York, NY, USA, 2018. View at: Google Scholar
  14. V. J. Cassol, “Evaluating and optimizing evacuation plans for crowd egress,” IEEE Computer Graphics and Applications, vol. 37, no. 4, pp. 60–71, 2017. View at: Google Scholar
  15. H. Cheng and X. Yang, “Emergency evacuation capacity of subway stations,” Procedia – Social and Behavioral Sciences, vol. 43, pp. 339–348, 2012. View at: Publisher Site | Google Scholar
  16. B. Haworth, M. Usman, G. Berseth, M. Khayatkhoei, M. Kapadia, and P. Faloutsos, “CODE: crowd-optimized design of environments,” Computer Animation and Virtual Worlds, vol. 28, no. 6, p. 1749, 2017. View at: Publisher Site | Google Scholar
  17. W. Lei, A. Li, and R. Gao, “Effect of varying two key parameters in simulating evacuation for a dormitory in China,” Physica A: Statistical Mechanics and Its Applications, vol. 392, no. 1, pp. 79–88, 2013. View at: Publisher Site | Google Scholar
  18. E. Ronchi, M. Kinateder, M. Müller et al., “Evacuation travel paths in virtual reality experiments for tunnel safety analysis,” Fire Safety Journal, vol. 71, pp. 257–267, 2015. View at: Publisher Site | Google Scholar
  19. Y. Hong, D. Li, Q. Wu, and H. Xu, “Dynamic route network planning problem for emergency evacuation in restricted-space scenarios,” Journal of Advanced Transportation, vol. 2018, Article ID 4295419, 13 pages, 2018. View at: Publisher Site | Google Scholar
  20. N. Shiwakoti, R. Tay, P. Stasinopoulos, and P. J. Woolley, “Likely behaviours of passengers under emergency evacuation in train station,” Safety Science, vol. 91, pp. 40–48, 2017. View at: Publisher Site | Google Scholar
  21. M. Haghani and M. Sarvi, “Pedestrian crowd tactical-level decision making during emergency evacuations,” Journal of Advanced Transportation, vol. 50, no. 8, pp. 1870–1895, 2016. View at: Publisher Site | Google Scholar
  22. L. Zhang, M. Liu, X. Wu, and S. M. Abou Rizk, “Simulation-based route planning for pedestrian evacuation in metro stations: a case study,” Automation in Construction, vol. 71, no. Part 2, pp. 430–442, 2016. View at: Publisher Site | Google Scholar
  23. J. Wang, M. Chen, B. Jin, J. Li, and Z. Wang, “Propagation characteristics of the pedestrian shockwave in dense crowd: experiment and simulation,” International Journal of Disaster Risk Reduction, vol. 40, p. 101287, 2019. View at: Google Scholar
  24. S. Cao, X. Liu, M. Chraibi, P. Zhang, and W. Song, “Characteristics of pedestrian’s evacuation in a room under invisible conditions,” International Journal of Disaster Risk Reduction, vol. 41, p. 101295, 2019. View at: Publisher Site | Google Scholar
  25. J. Wang, B. Jin, J. Li, F. Chen, Z. Wang, and J. Sun, “Method for guiding crowd evacuation at exit: the buffer zone,” Safety Science, vol. 118, pp. 88–95, 2019. View at: Publisher Site | Google Scholar
  26. P. Geoerg, J. Schumann, S. Holl, and A. Hofmann, “The influence of wheelchair users on movement in a bottleneck and a corridor,” Journal of Advanced Transportation, vol. 2019, p. 9717208, 2019. View at: Publisher Site | Google Scholar
  27. Z. Wang, F. Jia, E. R. Galea, and J. H. Choi, “A forensic analysis of a fatal fire in an indoor shooting range using coupled fire and evacuation modelling tools,” Fire Safety Journal, vol. 91, pp. 892–900, 2017. View at: Publisher Site | Google Scholar
  28. X. Chen, H. Li, J. Miao, S. Jiang, and X. Jiang, “A multiagent-based model for pedestrian simulation in subway stations,” Simulation Modelling Practice and Theory, vol. 71, pp. 134–148, 2017. View at: Publisher Site | Google Scholar
  29. S. Chen, Y. Di, S. Liu, and B. Wang, “Modelling and analysis on emergency evacuation from metro stations,” Mathematical Problems in Engineering, vol. 2017, Article ID 2623684, 11 pages, 2017. View at: Publisher Site | Google Scholar
  30. K. Fridolf, D. Nilsson, and H. Frantzich, “Fire evacuation in underground transportation systems: a review of accidents and empirical research,” Fire Technology, vol. 49, no. 2, pp. 451–475, 2013. View at: Publisher Site | Google Scholar
  31. W. K. Chow and C. M. Y. Ng, “Waiting time in emergency evacuation of crowded public transport terminals,” Safety Science, vol. 46, no. 5, pp. 844–857, 2008. View at: Publisher Site | Google Scholar
  32. L. Hong, J. Gao, and W. Zhu, “Simulating emergency evacuation at metro stations: an approach based on thorough psychological analysis,” Transportation Letters, vol. 8, no. 2, pp. 113–120, 2016. View at: Publisher Site | Google Scholar
  33. M. Marzouk and B. Mohamed, “Integrated agent-based simulation and multi-criteria decision making approach for buildings evacuation evaluation,” Safety Science, vol. 112, pp. 57–65, 2019. View at: Google Scholar
  34. E. Ronchi, S. Arias, S. La Mendola, and N. Johansson, “A fire safety assessment approach for evacuation analysis in underground physics research facilities,” Fire Safety Journal, vol. 108, p. 102839, 2019. View at: Publisher Site | Google Scholar
  35. Q. Li, Y. Gao, L. Chen, and Z. Kang, “Emergency evacuation with incomplete information in the presence of obstacles,” Physica A: Statistical Mechanics and its Applications, vol. 533, p. 122068, 2019. View at: Publisher Site | Google Scholar
  36. J. Chen, J. Ma, and S. M. Lo, “Event-driven modeling of elevator assisted evacuation in ultra high-rise buildings,” Simulation Modelling Practice and Theory, vol. 74, pp. 99–116, 2017. View at: Publisher Site | Google Scholar
  37. L. Jin, M. Xiang, S. Chen, X. Zheng, R. Yao, and Y. Chen, “An orderly untangling model against arching effect in emergency evacuation based on equilibrium partition of crowd,” Discrete Dynamics in Nature and Society, vol. 2017, Article ID 2757939, 7 pages, 2017. View at: Publisher Site | Google Scholar
  38. Y. Liao, G. Liao, and S. Lo, “Influencing factor analysis of ultra-tall building elevator evacuation,” Procedia Engineering, vol. 71, pp. 583–590, 2014. View at: Publisher Site | Google Scholar
  39. M. Usman, D. Schaumann, B. Haworth, M. Kapadia, and P. Faloutsos, “Joint parametric modeling of buildings and crowds for human-centric simulation and analysis,” in International Conference on Computer-Aided Architectural Design Futures, pp. 279–294, Springer Nature Singapore Pvt Ltd., Singapore, 2019. View at: Publisher Site | Google Scholar
  40. G. Berseth, B. Haworth, M. Usman, and D. Schaumann, “Interactive architectural design with diverse solution exploration,” in IEEE Transactions on Visualization and Computer Graphics, IEEE, Osaka, Japan, 2019. View at: Google Scholar
  41. G. Wu and H. Huang, “Modeling the emergency evacuation of the high rise building based on the control volume model,” Safety Science, vol. 73, pp. 62–72, 2015. View at: Publisher Site | Google Scholar
  42. Chongqing rail transit Group, “The operation route of Chongqing rail transit,” http://www.cqmetro.cn/wwwroot_release/crtweb/, January 2, 2019. View at: Google Scholar
  43. Chongqing Rail Transit Group, Annual Report on Traffic Development in the Main City of Chongqing, Chongqing Rail Transit Group, Chongqing, China.
  44. A. Borshchev, Y. Karpov, and V. Kharitonov, “Distributed simulation of hybrid systems with AnyLogic and HLA,” Parallel Computing, vol. 18, no. 6, pp. 829–839, 2002. View at: Publisher Site | Google Scholar
  45. Y. G. Karpov, R. I. Ivanovski, N. I. Voropai, and D. B. Popov, “Hierarchical modeling of electric power system expansion by anyLogic simulation software,” in Power Tech, 2005, pp. 1–5, IEEE, Russia, 2005. View at: Publisher Site | Google Scholar
  46. G. I. Hawe, G. Coates, D. T. Wilson, and R. S. Crouch, “Agent-based simulation for large-scale emergency response: a survey of usage and implementation,” ACM Computing Surveys, vol. 45, no. 1, 2012. View at: Publisher Site | Google Scholar
  47. K. Zhu, Y. Yang, and Q. Shi, “Study on evacuation of pedestrians from a room with multi-obstacles considering the effect of aisles,” Simulation Modelling Practice and Theory, vol. 69, pp. 31–42, 2016. View at: Publisher Site | Google Scholar
  48. L. Tan, M. Hu, and H. Lin, “Agent-based simulation of building evacuation: combining human behavior with predictable spatial accessibility in a fire emergency,” Information Sciences, vol. 295, pp. 53–66, 2015. View at: Publisher Site | Google Scholar
  49. Q. Li, L. Song, G. F. List, Y. Deng, Z. Zhou, and P. Liu, “A new approach to understand metro operation safety by exploring metro operation hazard network (MOHN),” Safety Science, vol. 93, pp. 50–61, 2017. View at: Publisher Site | Google Scholar
  50. A. Nieto-Morote and F. Ruz-Vila, “A fuzzy approach to construction project risk assessment,” International Journal of Project Management, vol. 29, no. 2, pp. 220–231, 2011. View at: Publisher Site | Google Scholar
  51. J. Shi, A. Ren, and C. Chen, “Agent-based evacuation model of large public buildings under fire conditions,” Automation in Construction, vol. 18, no. 3, pp. 338–347, 2009. View at: Publisher Site | Google Scholar
  52. K. Imanaka and K. Yasui, Research on Pedestrian Walking Speed and Behavior in Crosswalk, Nihon University, Tokyo, Japan.
  53. R. W. Bohannon, “Comfortable and maximum walking speed of adults aged 20–79 years : reference values and determinants,” Age and Ageing, vol. 26, pp. 15–19, 1997. View at: Google Scholar
  54. Y. Matsui, M. Hitosugi, T. Doi, S. Oikawa, K. Takahashi, and K. Ando, “Features of pedestrian behavior in car-to-pedestrian contact situations in near-miss incidents in Japan,” Traffic Injury Prevention, vol. 14, pp. S58–S63, 2013. View at: Publisher Site | Google Scholar
  55. N. Shiwakoti, R. Tay, P. Stasinopoulos, and P. J. Woolley, “Likely behaviours of passengers under emergency evacuation in train station,” Safety Science, vol. 91, pp. 40–48, 2017. View at: Publisher Site | Google Scholar
  56. X. Z. Zheng, X. L. Xie, D. Tian, J. L. Zhou, and M. Zhang, “Analysis of the evacuation capacity of parallel double running stairs in different merging form based on simulation,” Discrete Dynamics in Nature and Society, vol. 2018, Article ID 3541420, 14 pages, 2018. View at: Publisher Site | Google Scholar
  57. Y. Li, H. Wang, C. Wang, and Y. Huang, “Personnel evacuation research of subway transfer station based on fire environment,” Procedia Engineering, vol. 205, pp. 431–437, 2017. View at: Publisher Site | Google Scholar
  58. M. Kostovasili and C. Antoniou, “Simulation-based evaluation of evacuation effectiveness using driving behavior sensitivity analysis,” Simulation Modelling Practice and Theory, vol. 70, pp. 135–148, 2017. View at: Publisher Site | Google Scholar
  59. T. Takahashi, “Assessments of introducing new technologies in disaster prevention planning,” Journal of Simulation Engineering, vol. 1, pp. 1–10, 2018. View at: Google Scholar
  60. P. Wang, P. Luh, S. Chang, and J. Sun, “Modeling and optimization of crowd guidance for building emergency evacuation,” in 4th IEEE Conference on Automation Science and Engineering, pp. 23–26, IEEE, Washington DC, USA, 2008. View at: Publisher Site | Google Scholar

Copyright © 2020 Hui Xu et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.


More related articles

 PDF Download Citation Citation
 Download other formatsMore
 Order printed copiesOrder
Views2269
Downloads949
Citations

Related articles

Article of the Year Award: Outstanding research contributions of 2020, as selected by our Chief Editors. Read the winning articles.