Pedestrian and Evacuation Dynamics

For the new version of the Dutch design guidelines for buildings, a threshold value for the capacity of emergency doors needs to be set. Innovative large-scale laboratory experiments have been performed to investigate the capacity of emergency doors during evacuation conditions. This paper focuses in particular on the relation between capacity and the independent variables population composition and stress level.

It turns out that the average observed capacities are for all widths lowest for the lowest stress level and highest for the highest stress level. The population with a greater part of children has the highest capacity (on average 3.31 P/m/s). This is mainly due to the smaller physical size of children compared to adults and elderly, which makes it possible that more children can pass a door at the same time than adults. The lowest capacity (on average 2.02 P/m/s) is found for the experiment with 5% disabled participants.

The time that it takes an occupant population to reach safety when descending a stairwell during building evacuations is typically described by measureable engineering variables such as stairwell geometry, speed, density, and pre-evacuation delay. In turn, engineering models of building evacuation use these variables to predict the performance of egress systems for building design, emergency planning, or event reconstruction. As part of a program to better understand occupant movement and behavior during building emergencies, the Building and Fire Research Laboratory at the National Institute of Standards and Technology (NIST) has been collecting stairwell movement data during fire drill evacuations of office buildings. These data collections are intended to provide a better understanding of this principal building egress feature and develop a technical foundation for future codes and standards requirements. To date, NIST has collected fire drill evacuation data in eight office building occupancies ranging from six to 62 stories in height that have included a range of stairwell widths and occupant densities.

While average movement speeds in the current study of 0.48 m/s ± 0.16 m/s are observed to be quite similar to the range of literature values, local movement speeds as occupants traverse down the stairwell are seen to vary widely within a given stairwell, ranging from 0.056 m/s to 1.7 m/s.

It is very important to predict human behavior in fire and to take measures for evacuation time however it is difficult to predict individual responses and dangerous to perform a field experiment. According to the Final Report on the Collapse of the World Trade Center Towers from the NIST, 2005, it took an average of 48 seconds for evacuees to descend to the floor below. It was more than twice the time it would ordinary take. The evacuees needed to be rescued more than once during their descent due to physical fatigue.

The evacuation behavior of evacuees that have descended staircases repeatedly causes physical fatigue or dizziness which obstructs evacuation traffic leading to evacuation congestion and delay. Evacuation congestion is amplified in high-rise buildings due to a decline in physical strength of evacuees caused by the inordinate length of evacuation lines and the confluence of new evacuees at each floor.

In this paper, a field evacuation experiment with 351 participants was performed. Through the statistical analysis of the evacuation behaviors recorded by CCTVs, the probability of calculating the evacuation congestion from each floor was estimated and occurrence patterns relating to time and vertical space were analyzed.

The availability of credible, accurate and comprehensive human egress data is critical for responsible engineering. However, human egress data is particularly complex and susceptible to omissions and severe limitations. Currently, much of the data available is incompletely described, inconsistently represented or difficult to access, making the misuse of data more likely. This article outlines the issues that can arise from the compromised representation of data and associated background information. A Data Portal design is described that should help existing data-sets be described more comprehensively, and aid future data collection efforts. As such, the portal should aid in the development of more accurate theories and in more responsible engineering activities.

Design of escape routes by human performance data depends on flow rates given by e.g. the maximum of the flow-density relation. Although there are many different studies about this relation, even so for the qualitative shape there is still no agreement. To enhance the empirical database we performed experiments with up to 107 test persons under laboratory conditions to study local densities, speeds and flow in planar corridors. Due to new methods of digital image processing our analysis is based on data (trajectories) of very high accuracy. In this contribution we present an extract of our current studies concerning speed, density and flow in planar corridors. We also show effects on results by different measurement procedures. We used different sizes of measurement areas to determine the influence of this parameter. In particular the density where speed reaches zero due to overcrowding is sensitive to this variation.

In pedestrian evacuations bottlenecks can be a crucial factor influencing the evacuation time. The main question involves the design of bottlenecks to enable unhindered pedestrian flow in order to optimize evacuation times. For better understanding of this problem, a set of experiments with pedestrians in different bottleneck-scenarios has been performed. The results enlarge the database and allow the testing of the basic assumptions of performance based egress design.

Within experimental series of evacuation trials the main focus is on two subjects of pedestrians flow dynamics: The fundamental diagram and bottleneck. In the following we present the bottleneck trials. The pedestrian experiments have been performed in well-controlled laboratory conditions. Special attention was given to the comparability of different experimental series because both variations of width and length of the bottleneck have an influence. Formation of lanes inside the bottleneck becomes less pronounced for wide bottlenecks, even for narrow bottlenecks the numbers of lanes don’t interfere with the linear relations between flow and bottleneck width. Based on the length of the bottleneck the behavior of pedestrians is different, going through a tunnel or a door. Short bottlenecks show strong non-stationarity.

Online simulation coupled with real-time measurements of pedestrians in public buildings is a novel application which can be used to increase the security and safety of pedestrians within those buildings. To receive realistic forecasts it is necessary to update the simulation constantly against reality. The real-time video analysis can thereby support the simulation with the necessary data. This proposed system is largely supported by the state observer of the control theory.

Within a laboratory one room model pedestrian behavior is monitored and analyzed by video cameras. Evolving data is subsequently provided for the simulation where flow rates are additionally recorded. Thus, two resulting passenger flow rates are available which can be compared by the simulation controller. The actuators, receiving information from the controller, can afterwards trigger an appropriate action. Two different actuators have been realized in the model in order to bring the simulation passenger flow closer to the observed passenger rates: the velocity controller adjusts the walking speed of the passengers and the flow generator actuator has the ability to match the passenger generation rate.

Results show that the simulation passenger flow curve converges to the real passenger flow. As expected, the simulation curve follows the real passenger rate with a certain delay. Nevertheless, the simulation model appears to reflect the behavior of the persons in an appropriate way. Further investigations will show which additional instruments can be used to refine the simulation actor behavior.

Effects of stair gradients and obesity on walking speeds on stairs were empirically investigated. The participants included a group of elderly people (n=18) and a group of young people (n=15). They were asked to ascend/descend four staircases with different gradients, as well as to walk on a flat surface, at both normal and fast speeds. The study found the effects of stair gradients. When walking speeds on stairs are estimated, the gradient should be taken into account. In contrast, the study found no effect of overweight (or moderate obesity) on speeds.

Many devices exist which are designed to enable individuals with disabilities to travel between floors via staircases. The range of devices includes wheelchairs that descend and ascend stairs, platforms and chairs installed on the staircase, wheelchair carriers which transport the individual while occupying a wheelchair, and evacuation chairs used during emergencies.

A summary of commercially-available devices is presented, categorized by design type and usage. Specifications are provided for characteristics such as overall size, device weight, travel speed.

A summary of existing international standards and testing methods which apply to these devices is presented. A description of standards development work in the area of emergency evacuation chairs is also included.

WenChuan Earthquake took place on May 12, 2008 in China. 69,227 were confirmed dead, but Shangzao junior high school was a miracle. There was no death or injured. It took over two thousand students and teachers only one minute and thirty-six seconds to evacuate safely from the two 5-floor buildings. One of the miracle resulted from their continuous evacuating training every semester. The paper summarized the evacuating experience of Shangzao junior high school firstly. Other trainings from high school to elementary school were analyzed after the earthquake. Be short of the guidance and organizations, there were some problems in the trainings. Speed should be focused as the No.1. Order is the most important especially in elementary school. Students should be told to run downstairs in two lines if the stairs were not wide enough. It was dangerous for students to put books on their heads while they ran down, because it led to tumble easily.

Evacuating people with reduced mobility (PRM) from multi-storey buildings can be a difficult task. A number of commercially available devices can be used to assist in moving the PRM to the ground however, there is little consistent data quantifying the relative performance of these devices. In this paper four commonly used assist devices, the Evac+Chair, Carry-Chair, Stretcher and Drag Mattress are used in a series of 32 evacuation trials designed to assess their performance. The trials involve moving a PRM from a wheelchair to the device, moving the PRM along a long corridor to a stair and then down 11 floors to the ground. The performance of the devices is then assessed in terms of travel speed on the flat and stairs, number of handlers required to operate and ease of overtaking by other stair users.

The purpose of this study is to deliver new data and to bring attention to the subject of evacuation of children. Evacuation characteristics such as flow, densities and walking speeds are in focus. Current literature on evacuation is based mostly on studies on adults. Ten Danish daycare centers participated in full scale evacuation experiments where two age groups 0-2 years and 3-6 years were analyzed separately. The overall findings were as follows. Flow through doors, walking speeds and densities were age-dependent and differed strongly from the data in existing literature. The results showed higher walking speeds in spiral stairs when the children were familiar with the evacuation path. Higher person densities and faster flow through doors were obtained among the children than found in the current literature on adults. Children in the younger age group were generally slower than the older children. The children walked slower in horizontal plane than adults, however they were keen to run during the evacuations, in the latter case their travel speed increased and exceeded the adults’. Since the evacuation characteristics of children differ in many ways from those of adults, nowadays models badly comprehend the evacuation behavior children.

Blended Wing Body (BWB) aircraft with around 1000 passengers and crew are being proposed by aircraft manufacturers. This type of aircraft configuration is radically different from conventional tube type passenger aircraft and so it is essential to explore issues related to both fire and evacuation for these configurations. Due to both the large size and the unusual nature of the cabin layouts, computer simulation provides the ideal method to explore these issues. In this paper we describe the application of both fire and evacuation simulation to BWB cabin configurations. The validity of the computer evacaution simulations is also explored through full-scale evacuation experiments.

In the past decade, significant effort has gone into the planning and execution of full-scale sea trials in an attempt to improve, calibrate and validate existing evacuation models for passenger ships. In September, 2009 two assembly exercises were conducted at sea onboard the RO-PAX ferry SuperSpeed 1 by team members of the EU-funded project SAFEGUARD. The exercises were conducted with passengers during routine sailings between the ports of Kristiansand, Norway and Hirtshals, Denmark. Between both trials, a total of 1,769 passengers were assembled, on day one, 902 passengers and on day two 867 passengers. As part of the data collection exercise, passenger response time data was collected – using video cameras – and passenger movement data was collected using a novel infrared (IR) based position logging system. This paper briefly describes the development and testing of the data acquisition system and briefly discusses preliminary results.

As traffic problems in major and medium cities of China are getting increasingly serious and the demands for public transit are rising dramatically, the construction of subways and light-rail transit systems is imperative in the construction of infrastructures in various major cities and is also the symbol of modernized metropolitans. The total length of subway to be built and completed in China by 2010 is 1200 km. However, the social influence would be great once the emergency occurs in metro. To realize and grasp the characteristics of evacuation behavior of passengers in metro emergency is the precondition for the operation enterprise to constitute effective planning of emergency evacuation and conduct the passengers’ evacuation to safe area reasonably and quickly. The discrepancy in economy, culture, population and number of operational subway routes in different cities are bound to result in different passenger evacuation characteristics in case of any subway emergency. In this article, by questionnaire survey method and statistical analysis method, an investigation was conducted on the safety awareness and safe evacuation behavior of subway passengers in three different regions in China, namely Beijing (BJ),Nanjing (NJ) and Guangzhou (GZ), based on which the characteristics of subway passengers evacuation behavior in these cities were analyzed.

Abstract In large stations of public transportation high crowd densities can lead to potential safety risks and to unnecessary delays. To assess the

actual

capacity of potential bottlenecks a deeper understanding on the route choice of pedestrians is of great importance. This paper investigates the factors that influence the route choice of pedestrians when facing a stair/escalator combination in a major Austrian train station. We employ random utility models on data sets of revealed and stated preferences. In particular we investigate the potential for heterogeneities in taste by employing

mixed logit models

. The results show that, first, crowding is an important factor for route choice, second, that the application of mixed logit models is appropriate and, last, that the use of both revealed and stated preference data adds valuable information.

In this contribution, we propose a new framework to evaluate pedestrian simulations by using Extended Range Telepresence. Telepresence is used as a virtual reality walking simulator, which provides the user with a realistic impression of being present and walking in a virtual environment that is much larger than the real physical environment, in which the user actually walks. The validation of the simulation is performed by comparing motion data of the telepresent user with simulated data at some points of the simulation. The use of haptic feedback from the simulation makes the framework suitable for training in emergency situations.

To date, there is little research concerning selection of exit paths in emergency and non-emergency situations. Whether an individual takes their cues from others, follows signs, uses some other unknown method or some combination of the three when navigating unknown structures needs clarification. We used a wearable eye-tracking system to record where subjects look when they are exposed to new environments to determine how they use their visual awareness to navigate, and whether they are using exit signage to do so. The system provides a video record of the subject’s point-of-view. This knowledge could ultimately help improve evacuation times and save lives during emergencies by allowing for better design of egress systems and/or models, particularly those where a subject has no prior knowledge of the layout of the structure, such as an office building, subway/rail station, or an airport.

There are many available measurement techniques for documenting people’s movement patterns, but not all of them are appropriate for unannounced evacuation experiments. An unannounced evacuation requires that participants are not influenced beforehand. In addition, experiments are performed in all types of buildings and low ceiling height is often a problem. This paper describes three techniques that can be used for unannounced evacuation experiments, namely (1) filming from above – cameras with wide angle lenses, (2) triangulation with two cameras, and (3) distance measurement with a laser scanner. The description is based on the results from a research study in which the three measurement techniques were tested and evaluated. The study also involved collection of data in unannounced evacuation experiments.

Studies on microscopic pedestrian requires large amounts of trajectory data from real-world pedestrian crowds. Such data collection, if done manually, needs tremendous effort and is very time consuming. Though many studies have asserted the possibility of automating this task using video cameras, we found that only a few have demonstrated good performance in very crowded situations or from a top-angled view scene. This paper deals with tracking pedestrian crowd under heavy occlusions from an angular scene.

Our automated tracking system consists of two modules that perform sequentially. The first module detects moving objects as blobs. The second module is a tracking system. We employ probability distribution from the detection of each pedestrian and use Bayesian update to track the next position.

The result of such tracking is a database of pedestrian trajectories over time and space. With certain prior information, we showed that the system can track a large number of people under occlusion and clutter scene.

In this article, the Pre-Warning (warning time) concept is developed: the time between an incident being noted by a member of staff (either directly or indirectly) and the raising of a general alarm. This represents the potential delay in staff response as they interpret their provision of cues and respond; a delay that may be procedural and/or cognitive. The theoretical basis of this concept is discussed, examples of incidents involving this factor are described, and data is examined from experiments and incidents to quantify the extent of the impact and the effect of this concept upon the ASET/RSET calculation. Examples of how Pre- Warning delay can influence RSET will be presented, along with a discussion of those procedures that are particularly susceptible to the delay and suggestions as to how this might be remedied.

This paper explores the concept of the Network Fundamental Diagram for two-dimensional pedestrian networks. In doing so, we investigate if the average performance of the network can be described as a function of the average density or accumulation of the network, in line with recent findings on discrete vehicular traffic networks. We show that this is indeed the case, by considering data from walking experiments and from simulation using a microscopic pedestrian model. It turns out that the shape of the diagram is determined by a number of factors, such as the shape and size of the area, its use and its function, and the composition of the pedestrian flow. Finally, we propose several applications of the diagram for pedestrian networks.

We present a real-time agent-based approach to modeling crowd behavior that is based on complementary psychological and engineering principles. The application focus is for developing realistic models that address not only the physical but also the psychological aspects of crowd behavior. Our approach to modeling the psychology of a crowd is based on the principle of emotional reflection. According to this principle, our emotions are evoked in response to our perception of other people’s emotions; hence emotions propagate through a crowd as a result of each person’s perception of other crowd member’s emotions as well as external factors. The emotional model is coupled with a movement model that is based on the social forces formulation, but with parameters modified to represent the current emotional state of each crowd member. We present the model along with results of how different emotional levels can affect the movement dynamics of crowds.

The purpose of the study was to explore how evacuation safety in historical buildings can be improved for people with various disabilities. Accounts of real experiences on how well evacuation routes in historical buildings are adapted to people with different types of impairments, as well as suggestions for safety enhancing measures, were collected by the use of focus group interviews. Some examples of problems reported when evacuating from historical buildings were, level differences on the way to and in evacuation routes, problems concerning orientation and problems with detecting the alarm signal. People had different needs depending on their type of impairment. This means that safety in evacuation routes must be given much more focus when improving accessibility in historical buildings. Measures must be taken to address the different needs of people with different types of impairments. The results can be used for the whole range of professions taking part in the project of planning refurbishments of historical buildings.

This paper presents results on the simulation of the evacuation of the city of Padang with approximately 1,000,000 inhabitants. The model used is MATSim (

www.matsim.org

). Three different strategies were applied: shortest path solution, user optimum, system optimum, together with a constraint that moves should reduce risk whenever possible. The introduction of the risk minimization increases the overall required safe egress time (RSET). The differences between the RSET for the three risk minimizing strategies are small. Further quantities used for the assessment of the evacuation are the formation of congestion and the individual RSETs (in comparison with the available SET).

We have considered the methods for improving efficiency in queuing systems by theoretical analysis and experiments. First, a queuing system which has plural service windows is studied. There are mainly two kinds of systems which are a parallel-type queuing system and a fork-type queuing system. Queuing theory is often used to analyze these queuing systems; however, it does not include the effect of walking distance from the head of the queue to service windows; thus, a walking-distance introduced queuing theory is investigated. By using this model, we have discovered that the suitable type of system changes according to the utilization of the system. We have also verified that when we keep one person waiting at each service window in the fork-type queuing system, the waiting time dramatically decreases. Secondly, we consider queuing systems in amusement parks. Plural people waiting in the queue move to get on a roller coaster at the same time; therefore, the efficiency of the system is improved by shortening the moving time. The result of the experiments indicates that the moving time decreases if people keep walking in the queue to start instantaneously.

This paper presents a stochastic model for emergency evacuation simulation such as for a subway station. The model, MCEVAC was developed using Monte Carlo and a macro-simulation technique; the model includes a number of random variables, such as occupant load, initial occupant load, and pre-evacuation time. A random event of a train fire including fire location, fire growth rate, and the effects of fire smoke, can be used to determine the accessibility of egress paths. As model output, a probability representation of the evacuation time can be obtained by taking into account all or selected random variables. For validation, the calculated exiting time shows agreement with that from a previous station evacuation study. For a hypothetical train fire, the predicted evacuation time consists of likely consequences of the required safe evacuation time in the form of a statistical distribution. While still under development, the model is expected to be used for performance-based fire safety design of transit stations and tunnels.

The buildings are usually divided into two categories in light of whether obstacles exist: the large space with no obstacles (C-type buildings) and the complex buildings with lots of obstacles (L-type buildings). In this paper, considering the aisle region attraction factor, we proposed a revised model which is suitable to simulate the evacuation process in the L-type buildings (such as classroom, theater, and stadium bleacher) based on our original cellular automata occupant evacuation model. Furthermore, the revised model is able to implement the function that transfers real time information of the occupant density at exit area to evacuees. At last, a case study of simulation evacuation process in a theatre was proposed.

Macroscopic models based on dynamic network flow theory are successfully applied to obtain lower bounds on real evacuation times [1]. The goal of our research is to tighten this lower bound and to make this macroscopic approach more realistic by taking into account clustering of evacuees - a sociological phenomenon observed in evacuation scenarios. A cluster of flow units in the network flow model represents families or cliques which tend not to move independently but as groups [2]. This fact is not covered by macroscopic approaches based on classical network flow theory. In this article, we take clustering into account and thus improve existing macroscopic network flow models. We focus on two different sizes of groups traversing the network, modeled as single flow units and cluster flow units the latter of which occupy

d

times as much capacity as single flow units. In this novel approach, we are given fixed amounts of single flow units and cluster flow units and minimize the time at which the last (single or cluster) unit reaches the target. We present an algorithm that gives a

2

-approximation for general networks and is optimal for the subclass of series-parallel networks.

Disasters analyses have brought the interest in improving response to emergencies. In many cases there is a need to understand how people within a built environment react in a fire building emergency. Architectural clues play a significant part in the decision making and time taken to evacuate due to an emergency. In this study, virtual reality simulated experiments have been constructed to study the human ability to stay oriented while moving through spaces, evaluating alternatives and making decisions. Two variable models were constructed and implemented in a simulated built environment, in which 100 subjects were tested as in evacuation experiment. Data analysis shows how these variables affect human behavior on each critical decision points. Results showed that some architectural clues such as: windows and colors are important factors in the process of determining and following a route in fire emergency evacuation.

The effects of fire products on pedestrians are introduced into a multigrid evacuation model, in which the space is discretized into smaller grids with the size of 0.1 m × 0.1 m and each pedestrian occupies 5×5 grid sites. The effects of fire products on pedestrians consist of two parts: the desired movement direction and the step frequency. The data of fire products obtained from the simulation results of a well-founded computational fluid dynamic (CFD) program, the Fire Dynamics Simulator (FDS) developed by the National Institute of Standards and Technology (NIST). With the multi-grid model, we investigated the walk routes of pedestrians in fires, and the evacuation time in scenarios with different fire intensities, pre-movement times or door widths.

The aim of the Hermes project is the development of an evacuation assistant to support security services in case of emergency in complex buildings and thus to improve safety at mass events. One goal of the project is to build models for pedestrian dynamics specifically designed for forecasting the emergency egress of large crowds faster than real-time using methods applied in high performance computing. We give an overview of the project and the modeling approaches used focusing on the runtime optimization and parallelization concepts.

On the assumption of advisory information concerning an imminent Tokai earthquake being officially announced, as a case example we developed a spatial-spot type agent-based simulation model for the Nagoya Station area, where several terminal stations are concentrated; in the model, agents played people on their way home, and such factors as the routes selected by agents and the spatial restrictions, e.g. passages, were taken into consideration. Basic on SOARS(Spot Oriented Agent Role Simulator)platform, we conducted a large-scale crowd simulation with 160,000 agents and analysis the change of space density in one hour to compare to the estimates given by Nagoya City, we analysis the result and also refer to this kind of project for implementing much higher functions.

The macroscopic network model for pedestrians described in this paper is part of the German research project Hermes. The purpose of Hermes is the development of an evacuation assistant. Key outputs of the network model are travel (evacuation) times, identification of bottlenecks, pointing out alternative escape routes and/or optimization of escape route usage. The model uses Dynamic User Equilibrium (DUE) to analyse the network. This way it is able to show pedestrian flow with transient congestion effects, leading to time-varying route choice during an evacuation. The results can be calculated before or during an emergency case. The calculation time is much shorter and the results can be shown quicker than with a common microsimulation movie or animation. But arguably a microsimulation can be more detailed and for absolute information like evacuation time more conventional.

Evacuation of a large number of residents is conceivable in case of urban fires following a large earthquake in a city. It is essential to implement effective evacuation measures for ensuring residents’ safety in the regional disaster prevention plan. We have been developing a simulation model based on a potential method for the evacuation behaviors of city residents in a post-earthquake fire. The model has been validated by comparing the evacuation behaviors of Tokyo City residents in the Kanto Earthquake Fire, where the distribution of fatalities calculated by this model was qualitatively similar to that reported by the survey of that time. In this paper, the evacuation behaviors of Tokyo City residents in the Kanto Earthquake Fire were simulated for validating the prediction function of this model in terms of the traveling trajectory of an evacuee.

As described in this paper, we analyzed the influence of the time necessary to begin coping behaviors on the damage caused by chemical terrorism. To calculate the damage of a chemical attack in a major rail station, our network model-based pedestrian simulator was applied with systems designed to predict hazards of indoor gas diffusion. Our network model is designed to conduct simulations much faster, taking less than few minutes for simulation with ten thousands of evacuators. Results of our analyses were used for the instruction of rail station managers in a tabletop exercise held by the Kitakyushu City Fire and Disaster Management Department.

In this paper we present a stochastic evacuation model specifically for high speed passenger trains. The proposed model is an object-oriented model in which passengers are represented using a cellular automata method and the train space by a fine network of 0.5 m x 0.5 m cells. The model is based on Monte Carlo methods in order to simulate the probability and effects of passengers’ actions and decisions during the evacuation process. The datasets used as default by the model are taken from video recordings of evacuation drills and virtual experiments conducted at the University of Cantabria. However, the flexibility of the model allows the user to modify this data. The results of this model are then compared with other validated evacuation models. The proposed model has a simple user interface and the results are given in real-time. This model could be a useful tool for evacuation management during real emergencies. The advantages of using a stochastic approach for modelling passengers’ behaviour in relation to a deterministic approach are discussed.

The present paper outlines different approaches to pedestrian modelling, classifying them according to their movement and collision avoidance algorithms. An interesting question, not yet fully explored, is whether the simulation output is markedly different between the two main agent-based approaches. The paper explores this question by considering the operational performance of two rail stations in the UK. A software tool with a grid-based approach to pedestrian movement (STEPS 4.0), and a tool with a continuous approach (Legion Studio 2006) have been used to evaluate the gate-line clearance and platform circulation of passengers with respect to proposed station upgrades and future passenger demand. The paper concludes by summarizing the advantages of each approach, in the light of the findings from these studies.

Egress models typically use one of three methods to represent the physical space in which the agents move, namely: coarse network, fine network or continuous. In this work, we present a novel approach to represent space, which we call the ‘Hybrid Spatial Discretisation’ (HSD), in which all three spatial representations can be utilised to represent the physical space of the geometry within a single integrated software tool. The aim of the HSD approach is to encompass the benefits of the three spatial representation methods and maximise computational efficiency while providing an optimal environment to represent the movement and interaction of agents.

We combine a macroscopic and a microscopic model of pedestrian dynamics with a bidirectional coupling technique to obtain realistic predictions for evacuation times. While the macroscopic model is derived from dynamic network flow theory, the microscopic model is based on a cellular automaton. Output from each model is fed into the other, thus establishing a control cycle. As a result, the gap between the evacuation times computed by both models is narrowed down: the microscopic approach benefits from route optimization resulting in lower evacuation times. The network flow approach is enriched by including data of microscopic pedestrian behavior, thus reducing the underestimation of evacuation times.

The development of evacuation models in the last three decades has mainly contributed to the assessment of occupant safety and evacuation procedures in a variety of building designs, under a range of environmental conditions. The effectiveness of such evaluation relies mainly on the models ability to reflect the detailed interactions between the occupant, building design, and environment. The purpose of this study is to present emergency evacuation modeling as a novel approach to layout designs and evacuation procedures. The approach is based on the development of a novel evacuation model that adjusts its outputs to evaluate a range of layout designs. The proposed evacuation model relies on the application of evolutionary computation techniques to assess the means of egress by evolving the location and number of exits needed to ensure occupants safety. The performance of the algorithms varies by occupant behavior. The study suggests that the algorithms have the potential to be implemented in more complex design problems. The study further suggests the need to validate the configurations found by the algorithms by conducting actual evacuation drills.

In this article, the optimization problem of evacuation of a building with many escape routes with consideration of dynamic changes in people – building-threat interactions is described. The model of system conveying information of the best way of evacuation with taking into account the people who will not adapt to the recommendations is proposed. On the base of chosen strategy of evacuation in every place of the building where direction of evacuation can be changed, the proposed system will give full information about the evacuation possibilities and recommends one of the routs. In the agent model, the individual characteristics affecting the decision-making process are been proposed. Self developed microscopic model of human behavior was used to simulate the evacuation problems. The representation of model environment uses a combination of two approaches: coarse network model and fine network model simultaneously. During the simulation, the main factor influencing the path of movement of individuals is the information supplied by the shared information system of the best way. In addition to a certain probability can decide to change the route of the evacuation.

In this study, a framework model, based on lattice gas model and Mean-Field Approximation model, is proposed to analyze pedestrian evacuation uncertainty. The model is focused on the probability that each grid is occupied by a pedestrian, but not the specific pedestrians. By calculating the evolution of occupancy probabilities, more information about the simulation uncertainty can be gotten. The model is presented in terms of a series of nonlinear equations and complete probability formula. In each time step, the probabilities that pedestrians exist on each site, as well as the transition probabilities to the neighboring sites, are calculated and updated using random sequential update rule. The pedestrian flow going outside a single-exit room is investigated numerically. The cumulative distribution and probability density distribution of the total evacuation time can be obtained by a single simulation using the model. In this case, the uncertainty and reliability of the simulation results can be easily analyzed and improve the calculation efficiency. In addition, the time dependent of mean flow rate and the effect of the width of exit on the total evacuation time are studied. The framework model can be extended using other Cellular Automaton model with different rules to analyze their uncertainty.

In this article, we propose a novel modeling approach – the sandwich approach – to deal with evacuation time bounds (ETB) - in which lower and upper bounds for the evacuation time are calculated. A provable lower bound is achieved by computing a quickest flow, using a dynamic network flow model, an upper bound is obtained via simulation using a cellular automaton model. Coherence between the macroscopic network flow and the microscopic simulation model will be discussed. In order to validate our theoretical results, we report on our practical experiences with the Betzenberg, the region containing the Fritz-Walter soccer stadium in Kaiserslautern, Germany.

This paper illustrates a formal tool for knowledge representation and management in the crowding area, in order to improve on the sharing of knowledge, data and information produced by different models and simulation tools. The presented approach exploits knowledge-based methods for acquisition and representation phases.

After a short discussion about crowd research area, we will present the methodology we have used to develop the tool. It is composed by an ontology and a set of fuzzy rules, which provided crowd classification according to a sociological theory previously formalized.

At last, in order to show how using the tool, a case study on a particular crowd scenario is proposed.

The article presents a new approach to crowd compressibility modeling in pedestrian evacuation. The model is based on Social Distances Model and implements a compressibility coefficient. The main purpose was to study specific flow of pedestrians through bottlenecks. Real data from two experiments were used to validate received results. Differences in compressibility parameters significantly influence pedestrian behavior and simulation scenarios. Higher values of compressibility coefficient lead to increased densities and flow of pedestrian stream.

People die or get injured at mass events when the crowd gets out of control. Urbanization and the increasing popularity of mass events, from soccer games to religious celebrations, enforce this trend. Thus, there is a strong need to find a better means to control crowd behavior. Here, simulation of pedestrian streams can be very helpful: Simulations allow to run through a number of scenarios in a critical situation and thereby to investigate adequate measures to improve security. In order to make realistic, reliable predictions, a model must be able to reproduce quantitatively the data, known from experiments. Therefore, automatic and fast calibration methods are needed that can easily adapt model parameters to different scenarios. Also, the model must be robust: Small changes in the crucial input parameters must not lead to large changes in the simulation outcome. In this paper we represent two methods to automize the calibration of pedestrian simulations. We then introduce a concept of robustness to compare the two methods. In particular, we propose a quantitative estimation of parameter quality and a method of parameter selection based on robustness criteria.

In this paper, two different pedestrian movement simulation models (a model of the social force type and a queuing network model) are compared with respect to their capability to predict individual walking times in a crossing area. Both models are calibrated using a trajectory data set and their relative performance on the estimation data set as well as on a separate validation data set is discussed. The social force type model is found to better predict the walking times as well as space usage in both in- and out-of sample comparison.

We discuss several aspects related to the validation and calibration of cellular automata based models of pedestrian dynamics. Empirical fundamental diagrams obtained in large-scale experiments are compared with simulations of the floor field model. Although this kind of macroscopic calibration gives insights into the relevant interactions that govern the collective behaviour of pedestrians, microscopic validation based on the trajectories is expected to be more reliable. As a simple scenario we consider trajectories of the motion of individual pedestrians around corners. It is found that besides interactions with walls inertia effects play an important role for the correct reproduction of the empirical trajectories.

Commercial software-tools for evacuation calculation are mostly closed-source (also called “proprietary”) software-tools. Contrary to open-source software tools users of closed-source software-tools have to trust the technical manual of the software to understand how the software works. For users it is important to know how the model they use works, because without this knowledge results calculated by the software-tool can be interpreted in a wrong way. We will present a methodology how to identify basics of evacuation modelling and how to interpret and understand calculated results in a better way. This methodology should give users a “better feeling” in a very short time for the model they use for evacuation analysis and calculation

The main topic of this paper is the analysis and modeling of stop-andgo waves, observable in experiments of single lane movement with pedestrians. The velocity density relation using measurements on a ‘microscopic’ scale shows the coexistence of two phases at one density. These data are used to calibrate and verify a spatially continuous model. Several criteria are chosen that a model has to satisfy: firstly we investigated the fundamental diagram (velocity versus density) using different measurement methods. Furthermore the trajectories are compared to the occurrence of stop-and-go waves qualitatively. Finally we checked the distribution of the velocities at fixed density against the experimental one. The adaptive velocity model introduced satisfies these criteria well.

This paper describes the characteristics of crowd flow passing through simple-shaped rooms and to validate an evacuation simulation model called “SimTread”. It presents experiments on crowd flow with 43 subjects and intends to quantify the aspects of crowd flow in simple rooms. Recently, for estimating evacuation of buildings on fire, several computer simulation models have been developed and applied. However, human evacuation data-sets for validating simulation are scarce. The results of this study show that pedestrian flow rate at the opening changes depending on the density of the space connected. Flow rate rises if the opening is connected to larger space, which is less dense and, in result, increases speed of pedestrians. For validating an evacuation simulator, evacuation data from actual building fires are too complex for proving the equivalence. Therefore, we carried out the experiments of crowed flow passing through simpleshaped rooms, and compared experimental data with our simulation results. There was a good agreement between the result of experiments and simulations. The differences were less than 10%.

Recently, evacuation analysis, which is one of the performance-based evacuation design methods, has been widely used to estimate the egress capacity of buildings and analyze evacuees’ characteristics. But almost all the models do not consider accurate evacuation scenarios because of the limitation of simulations, especially when setting up the staircase-length in building geometry mode. Therefore, this study selected a subject of a high-rise building with 351 invited participants, and conducted an experiment to investigate the evacuation from the 30th floor to the ground level and then classified overall 5,265 movement cases to calculate the participants' evacuation route. After the trial experiment, the paper calculated the precise average distance of participants' evacuation route in staircases considering each zone’s length and dimension and derived the numerical equations for inputting the parameter of staircase-length in evacuation models. And the verifications were also conducted by comparing simulations with the results of the experiment.

US model codes and building regulations are recognizing the provision of protected elevators for occupant self-evacuation after more than two decades of training people that elevators are unsafe in fires. This reversal will require that people can readily identify those elevators that are safe to use and be provided with information and reassurances during use that the system is functioning safely. Lengthy discussions on the interactions between the systems and users have resulted in requirements for visual, audible, and voice messaging systems and operational protocols designed to provide reliable, real-time information needed by users to make informed decisions. The discussions have further identified the need for public education to provide for effective use by infrequent visitors to buildings equipped with these systems.

The paper will discuss the approaches being developed to address these needs by a consortium of public and private organizations including the American Society of Mechanical Engineers (ASME), National Institute of Standards and Technology (NIST), National Elevator Industry Inc. (NEII), disability advocacy groups, and the fire alarm and model building code developers. There is an expectation that since a building’s elevators are used daily by the occupants, by keeping the system used in emergencies as close as possible to normal use, the provision of additional information on status and safety will represent sufficient reassurance to users.

Otion that the approaches will be effective. Due to the difficulties inherent in human testing, there is a need for the inclusion of these features into observational research being conducted through required evacuation drills. Since elevator use is not prohibited now for non-fire emergency egress, this may provide the opportunity to test public response to the approaches being contemplated. The paper will suggest ideas for such research being included in planned studies and as a part of building commissioning.

This paper presents an overview of human factors data collected via an online survey related to the use of lifts (elevators) and stairs during both circulation and evacuation scenarios. Survey participants were presented with a series of hypothetical situations and asked how they would behave. The survey was split into two broad sections, the first dealing with normal circulation usage of lifts/stairs and the second dealing with evacuation usage of lifts/stairs. Detailed demographic information about each participant was also collected. In total some 468 people from 23 countries completed the survey. An overview of the survey and initial results are presented in this paper.

In emergency situations, the practice has been to return elevators to the exit discharge level, and then shut down. After that, the elevators are not available for the building occupants until the emergency is over. In this paper, we study the use of the elevators in an emergency evacuation. We first introduce the theoretical egress time and the handling capacity calculation for an elevator group, which is based on floor-by-floor evacuation. We also describe a specialized elevator evacuation algorithm, which automatically, with or without landing call information, dispatches elevators to the occupied floors and shuttles passengers to the rescue level. The algorithm serves floors in a priority order and detects floor occupancy automatically. We compare this algorithm to a normal algorithm and to two staircases in evacuation. For that purpose, we run simulations of test buildings with realistic transport arrangements and obtain performance measures such as crowding levels of the lobbies, passenger service times and total evacuation time. On this basis, we propose the best algorithm for different types of emergencies.

This paper reports experiments on the reaction of pedestrians to the sudden stop and restarting of an escalator to explore the applicability of escalators moving toward safer side. Deceleration rate at the stop and acceleration at the restarting were controlled so that optimum operation to keep the safety of pedestrians were sought. 32 experiments were conducted on each of 10 subjects(6 females and 4 males) by changing the moving/still mode of the escalator, moving direction of escalator(upward/downward), deceleration and acceleration rates, walking/still standing modes of the subjects, and burden conditions. The experiments indicate restarting generally safer than stopping, downward operation generally safer than upward one, and significance of the influence of the burden condition to the safety of pedestrians at emergency stopping and restarting. The difference whether the pedestrian walks or stands still did not cause significant difference in the pedestrians’ safety. The experiments also indicate significant effectiveness of the reduction of the deceleration rate at emergency stopping from the current 0.61 m/s

2

to 0.43 m/s

2

and the acceleration rate at emergency restarting from the current 0.17 m/s

2

to 0.10 m/s

2

for the improvement of the pedestrians’ safety.

This paper presents a study at the Manchester United Football Stadiumat Old Trafford, UK. The purpose of the study was to assess the safety risks associated with spectators standing in an all-seated stadium. Information gathering was a key part of the study; this paper looks at the information gathering exercises undertaken and how they contributed to the risk assessment. It also presents some of the key study findings. It may not be immediately apparent that standing in a stadium that is not overcrowded could be an issue. But there are in fact a variety of movements and behavioral factors involved when people stand. It is these movements and behaviors and their interactions with the physical environment that was our focus of investigation. Overall, the study has found that the risk of standing varies depending on the situation and the physical design of the seated area. There are a number of risk factors involved; we believe that many of them can be controlled through stadium design and crowd management

Human behaviour in building evacuation and use of elevators has been studied more extensively in recent years. There is also a move to explore concepts of “protect-in-place”, thus avoiding evacuation in some circumstances. This raises the matter of decision making as to whether to “stay or go”. In the fields of natural hazards, such as bushfires and floods, this decision to “stay or go” becomes equally important for safety. A comparison of literature sources on human behaviour and decision making in buildings with recent research on bushfires and floods highlights some common factors critical to decision making as to evacuation or not. These factors include emergency preparedness, situation awareness and trusted information systems. This paper suggests that greater understanding of group behaviour and socio-cultural differences is required if more effective emergency management is to be achieved.

Survey responses from occupants involved in a 32-story high-rise building evacuation during a fire were collected and analyzed to study the preevacuation period. Multiple regression models were used to test whether specific occupant, building, and environmental factors predicted pre-evacuation times. This study found that the actions taken by occupants during this period, i.e., waiting, helping, and preparation actions, was a main factor that significantly increased pre-evacuation time. In addition, although the action of seeking information did not significantly influence pre-evacuation time, interaction effects were found among certain groups of occupants who sought information. Of those who sought additional information, older adults took less time seeking information (per action) than younger adults, occupants with disabilities took more time seeking information (per action) than occupants without disabilities, and occupants who perceived risk/danger took less time seeking information (per action) than those who did not perceive risk. This study also found that the initial location of the occupant (floor) also significantly influenced pre-evacuation time, likely due to the information that occupants received on these floors. The effects of these factors on preevacuation times are quantified by stating how much pre-evacuation time was required for each factor (i.e., action taken or floor location).

Twenty–twenty four percent (20-24 %) of the world population are people with disabilities. This population has special access needs for performing activities of daily living. This is especially important in case of an emergency. The evacuation concept for a building must take people with disabilities into consideration by means of including specific parameters as an integral aspect in the evacuation simulation model.

This paper presents the accessibility parameters to be considered in evacuation simulations. Examples of tools used in the two disciplines (evacuation and accessibility planning) are provided for illustration: The PedGo evacuation simulation package [Kluepfel, 2003], a model which does only implicitly take into account disabilities; and a decision support system for evaluating accessibility of facilities [Bendel, 2006] illustrates the accessibility approach to facility planning.

Recommendations for integration of accessibility and evacuation analysis summarize the paper.

Transit authorities globally are developing their networks to make them more accessible for Persons with Reduced Mobility (PRM’s) whilst demographic and lifestyle trends suggest that the proportion of PRM’s within the population who are likely to make use of these newly accessible networks will increase over time. In the UK, London Underground has developed a modeling framework that incorporates PRM movement into pedestrian micro-simulation models. Network Rail, the UK mainline station operator has also been undertaking PRM surveys to determine the characteristics of PRM demand at mainline stations and this data validates the approach adopted by LUL which requires a location specific understanding of PRM populations, incorporates spatial and temporal variation, and is coupled to location specific characteristics (for example, Network Rail long distance passengers are less familiar with their station environment than commuters and require more information, signage and management. They are more likely to travel in groups and be encumbered).

Anecdotal evidence from studies completed in London suggests modeling output is sensitive to the volume of PRM’s modeled. To date, no detailed reviews have been undertaken to assess this sensitivity. This paper reviews the PRM modeling methodology used in London and assesses its applicability in the America’s and elsewhere. The relationship between maximum flow rates and the average speed and footprint of entities is tested through the use of Legion micro-simulation models and described through the application of multi-variable linear regression to the output. Impacts of modeling PRM’s are described and implications for Space Planning Performance Metrics and design recommendations are reviewed. 714 D. A. Fisher and A. Jenkins

Environmental uncertainties and the complexities of human behavior create challenges in the design of adaptive crowd management systems. The authors suggest a computer-assisted management system based on principles of cognitive systems engineering, including distributed decision-making agents with variable goals, and feedback loops that facilitate adaptation. A computer-based tactical crowd advisor uses a behavioral model to suggest actions to the crowd manager based on the behavioral model’s predictions of crowd movement. The crowd manager instructs security personnel and pedestrians, and observes the results, thus completing a feedback loop. An agent-based simulation is used to generate behavioral input so that the expected effects of physical and procedural features of alternative proposed systems. The basic principles of the cognitive systems approach adopted herein are illustrated using examples from an investigatory report of the Presidential Inauguration in Washington D.C. in January of 2009.

EVA is the acronym for an interdisciplinary research project addressing risk management of major events. The focus here is on the behavioral aspects of crowd movement at annual fairs, festivals, large parades or sport events. As a first step, video recordings at accentuated places are taken and analyzed. The respective findings are next included into the microscopic evacuation model ASERI developed by IST GmbH. The so expanded model is now in the process of validation against available empirical data and will be used to investigate crowd movement in normal and perturbed cases including emergency scenarios. Emphasis is put on social interactions and collective effects like group formation and persistence and the interaction of crowd flow with technical provisions and organizational measures. Ultimate goal of the EVA project is the development of a guideline on the evacuation of large scale event spaces in cooperation with German municipal fire brigades.

This paper presents the developed concept of management and control to the pedestrian flow movements at Jamarat in the regular annual pilgrimage to Makkah (Hajj) season in Saudi Arabia. Every year, 3 to 4 millions of pilgrims perform their rituals in the course of extensively high restrictions, in the midst of a climax of limited/narrow space and time constraint. The Jamarat where pilgrims gather to perform a ritual stoning of pillars symbolizing the devil as part of the Hajj. The new Jamarat leveled building replaced the old ones. The project objective is to prevent crowd panic and to minimize the risk of crowd disasters. Management and control of pedestrian group movement to and/or inside Jamarat leveled building and area, using new experimental knowledge methodologies observed from the science of crowd dynamics, throughout anticipation and analysis of the pilgrim flow from low crowd density to extremely high crowd density, accompanied, attended, and escorted with an insider real-time-life video- scrutiny/observation/analysis.

The current NFPA Life Safety Code and International Building Code have regulations of stairwell widths for high-rise buildings that are primarily based on a linear relationship of 7.6 mm of stair width per person based on the highest occupant load from a floor expected in the stairwell. This linear relationship between stairwell width and the speed and flow of occupants as they egress the building will be examined using data from four high-rise building evacuation drills that experience high density of occupants during egress and was collected by the National Institute of Standards and Technology.

For a proper understanding and modeling of pedestrian dynamics reliable empirical data are necessary for analysis and verification. Therefore we have performed a series of experiments with a large number of persons. For the time-efficient automatic extraction of accurate planar pedestrian trajectories we developed the program PeTrack. We now have extended the software to stereo recordings, which allows a direct height measurement without additional markers and to extract trajectories on stairs.

In this paper we present a generic methodology for calibrating walking models and we discuss some results of its application. This methodology aims to describe all aspects of the parameter estimation process of walker models, while making it applicable to the calibration of any walker model. The base of the methodology is the simultaneous calibration of several aspects of pedestrian traffic to improve the robustness and generic application of the model. Applying the methodology to the Nomad model showed that combining different flow configurations indeed improved substantially the quality and significance of the parameters.

New large-scale laboratory experiments have been performed to investigate the capacity of emergency doors during evacuation conditions. Varying doorway widths showed that only the experiment with the widest doorway (275 cm) resulted in a capacity lower than the capacity found in previous experiments (2.25 P/m/s). However, this was not due to the wide door, but to a population composition with less children.

Pedestrian route choice is a complex, situation- and populationdependent issue. In this contribution an example is presented, where pedestrians can choose among two seemingly very similar alternatives. The choice ratio is not even close to being balanced, but almost all pedestrians choose the same alternative. A number of possible causes for this are given.

This study looks at the possibility of exploiting sensor-linked modeling to provide information to occupants on the available routes within the building and help influence their final egress path to support efficient evacuation. In general, it is intended that the system will able to influence the occupants’ egress route choices, even before they have decided on their final exit choice. The effectiveness and reliability of the system will be tested through a series of comparisons obtained by modeling evacuations using K-CRISP (Koo [5]) in conjunction with live data that will be gathered during project M*E*T*R*O (Nilsson [1]).

This paper describes the application of Mutual Information to the detection of crush in a well-established model of pedestrian evacuation. We show that Mutual Information offers a computationally low-cost alternative to "expensive" physical force calculations for the detection of crush in evacuation simulations.

The authors present an application of agent based simulation in evacuation planning and emergency response. The Evacuation Planning Tool (EPT) developed by Regal Decision Systems Inc for DHS is a PC based software application that consists of a 3D Editor, an agent based model, a 3D animation component and a reporting module. The EPT provides users with the ability to simulate and analyze a wide variety of evacuation scenarios. The EPT simulates movements of crowds and provides users with the ability to control security teams at a command level. Users can assess civilian responses, facility operations and security options related to a variety of disasters and threats in order to analyze and formulate evacuation strategies. Entities and groups of entities are created with decision branching intelligence and dynamic selection of destinations, routes, and behaviors. The behaviors are consistent with environmental influences such as threats, physical barriers, encumbrances, and commands.

We present a new method for modeling counterflow situations in crowds. Agents, describing individual pedestrians, are set to avoid the moving directions where there is counterflow and prefer the directions with forward flow. In dense counterflow situations, people tend to move shoulder first to occupy less space in the moving direction. If the elliptical cross-section of a human body is considered in a crowd model, the rotational positions in which the agents move affect the counterflow. In our model, agents try to rotate their bodies in certain counterflow situations to move shoulder first. The model is implemented in the FDS+Evac simulation software. Test simulations show that it is able to create rather realistic simulations of counterflow.

This study attempted to re-create the impact of evacuation obstacles such as moved or tipped furniture and scattered goods on people's evacuation behavior in a building at the time of an earthquake in an evacuation simulation. We conducted an earthquake evacuation simulation of the time required to complete evacuation and evacuation behavior with simulation conditions based on the information obtained through the interviews with representatives of four consumer facilities which actually suffered damage from an earthquake, focusing on people's behavior both when the building was shaking and after the earthquake, the reaction of staff, damage inside the building and evacuation behavior in order to seek the impact of damage inside the building, assuming the tipped or moved furniture and scattered goods would partially block an evacuation route and thus cause a reduction in walking speed. As a result, a phenomenon that people are crowded into sales spaces and passageways causing a traffic jam was re-created on the assumption that evacuation obstacles such as furniture and goods would slow down walking speed.

In agent-based pedestrian circulation models, the simulation of the pedestrian-environment interaction is mostly achieved by imposing on each agent a predefined list of goal locations which the agent visits in turn. However, in reality human behaviour in complex environments is highly dynamic and fixed plans are often changed and adjusted according to emergent conditions and the person’s individual interpretation of these events, in particular the amount of time available to achieve all the desired tasks. In this paper we present a prototype emotion model implemented within the buildingEXODUS evacuation and pedestrian dynamics software which enables simulated agents to react to perceived time pressures by modifying their behaviour. The model is demonstrated using a circulation scenario within a rail terminal.

Density of the person in a classroom is one of the important factors to evaluate building evacuation safety regulation if being fire a building. It provides the number of persons in the classroom for the unit area in according to building purposes on Code or Standard in the evacuation overseas countries of USA or UK and in Japan also, institute of Fire organizes a research committee and investigates the number of persons in the classroom for every service under cooperation of Tokyo Fire Department and Construction Companies, and prescribes to consider them for design a building.

However, it is approved only a specified criterion as a law without a criterion in according to building purposes in Korea, and then it has limitation to evaluate whether the person in a classroom has evacuation stability in multiple buildings and some of part might not be suitable with the condition in Korea, even to fit the density criterion of the person in a classroom of overseas simply, because it might be difference a structural characteristic for using and composing for every country even though buildings have the same purposes.

For that, this study presented density of the person in a classroom after investigating survey research of buildings that has high risk to make injuries to persons if being fire in a large shopping mall, multiple buildings and a large multiple theater for setting up the density criterion of the person in a classroom that is suitable to the condition in Korea.

Korean railroad company is continuously improving the service for the disabled. They also install and reform many facilities for them. However, It is not considered the importance of egress when fire occurred in a cabin Now, Fire hazard analysis and risk assessment are introduced into domestic regulations for improving fire safety performance of train car and railroad facilities. Considering the fact that fire safety evaluation process doesn't fully include disabled person, fundamental egress data for them is very important for life safety aspect. In this paper, Egress experiments and analysis were conducted to measure the movement speed of the disabled with Mugunghwa and Saemaul.

This paper deals with a problem of time scaling and validation of a mathematical model of a pedestrian flow. We focus on stochastic cellular automata approach. What kind of tests should be applied to say that model ``works''? In this paper some our tests and time scaling observations are presented.

The recent developments of the evacuation programme FDS+Evac are presented including a counterflow model for continuum crowd modelling. The major new additions are described and some verification and validation results are presented including an analysis of specific flows through doors and corridors. The presented additions are implemented in the latest version of FDS+Evac. This version is embedded in the computational fluid dynamics based fire simulation programme Fire Dynamics Simulator, which is freely obtainable from the FDS web page

http://fire.nist.gov/fds/

.

This contribution reports on the research project SKRIBT and some of its results. An evacuation simulation based on VISSIM's pedestrian dynamics simulation was developed, that - with high time resolution - integrates results from studies on behavior in stress and crisis situations, results from CFD models for e.g. fire dynamics simulations, and considers visibility of signage and — adding a psychological model — its cognition. A crucial issue is the cognition of smoke or fire by the occupant and his / her resulting spontaneous or deliberate reaction to this episode.

Standard definitions of the density exhibit large fluctuations when the size of the measurement area is comparable with the size of a pedestrian. An alternative measurement method exists where a personal space, calculated through the Voronoi diagram, is assigned to each pedestrian. In this contribution this method is applied to an experiment studying motion through a bottleneck and the reduced fluctuations demonstrated. The integrated density also permits examination on much smaller spatial scales than the standard definition, the insights into the pedestrian motion this provides are discussed.

We introduce an intelligent lighting system for dynamic pedestrian guiding. Its benefits in emergency situations are demonstrated through a simulated evacuation scenario. A pilot system for passenger flow optimization is presented as an example of a non-emergency application. The system is useful for implementing solutions for pedestrian guiding found through modeling and simulations.

Requirements for ever larger, architecturally complex, structurally challenging and economically efficient buildings, but with a constantly growing number of users for buildings creates a situation that in the event of an emergency scenarios (fire) can cause significant complications for rescuing people. Ensuring the safe evacuation of persons from buildings is a fundamental requirement in the design of fire safety works. Current standardization requirements are difficult to enforce in the design of existing multifunctional buildings. The paper presents the results of people-evacuation modeling in two buildings in Slovakia, which were only recently completed. The paper presents a comparison of calculation results with standardization values. Part of the contribution is also a discussion of the results and their interpretation. The conclusion allows us to use the results of the evacuation modeling after changing the actual requirements.

Both experimental and numerical studies indicate local interaction among pedestrians plays an important role in determining the self-organization of pedestrian traffic. However, little is know to the form of the interaction. We explore the fundamental interaction between pedestrians by performing controlled experiments in a long narrow channel where pedestrians were asked to first walking freely in the channel, then to evade a standing still pedestrian in the middle of the channel. Trajectories of these pedestrians extracted from the video-recordings of the experiments were then used to analyze detailed microscopic moving features. Averaged change of speed as well as the change of direction of these pedestrians was then calculated. An angular and metric distance based interaction map reviewing the laws ruling the behavioral changes when a pedestrian interacts with others was detailed. This study provides further insights to the fundamental individual level interaction between pedestrians for understanding the pedestrian dynamic and is expected to be helpful in evacuation analysis.

In the past few years, the world has witnessed severe fires in nightclubs, which have caused many human losses. This paper promotes a methodology which uses evacuation modeling analysis for improving the fire safety strategies (FSS) in nightclubs. The methodology is presented through a study case which represents a nightclub scenario. The results have shown that the use of evacuation modeling analysis can help immensely fire protection engineers to develop their FSS reports, incorporating the human behavior factor; rather than only considering the standard safety factors, such as travel distance. The same methodology can be also applied for fire safety management, which is another key-issue to be properly addressed for fire safety purposes in nightclubs. It is expected that this paper can bring some additional light to the way emergency evacuations are planned and addressed is in nightclubs.

At major transit terminals large, volumes of people, intricate operational procedures, and complex built environments present significant challenges to effective pedestrian facility design. Transbay Terminal in San Francisco is a prime example of layered complexity. It is a multi-modal transit terminal designed to serve commuter rail, commuter bus, local bus, and eventually high speed rail passengers in downtown San Francisco. With the existing facility already near capacity and new transit modes being planned it is essential that the terminal facility is upgraded. This paper will present a new crowd simulation technology called MassMotion and describe how this toolset was applied to inform the design of a new Transbay terminal.

This paper discusses the development of object oriented, agent based facility simulation tools geared to examining not just physical movement performance, but operational interactions from overarching facility command and control to individual human reactions as experienced during both normative and extreme scenarios.

The paper examines how the next generation of facility modeling will make use of agents moving within 3-dimensional space physically interacting with each other using onboard stochastic rules that define their actions and movement geometries. Human agents will have both a physical kinematic presence as well as visual and auditory capabilities to apprehend their environment.

The suite of agents involved in the “facility” model will expand from simply representing human actors and simple mechanisms such as escalators and elevators to include the “equipment” pedestrians and facility operators use, both large and small: from cell phones, to wheelie luggage, to signage, to kiosks and stores and their contents, to trains, buses and cars, to their controlling signaling and their centralized command and control centers, to the air and smoke they breathe.

Expansion of the physical components that can independently interact will facilitate much more realistic next generation analyses of facility operations under exceptional conditions including the disorienting effects of fire, smoke, noise, poor lighting, miscommunication, equipment failure, etc. Such “deep” models will produce new understandings of the interaction of humans in the physical environment as distorted under extreme event conditions.

The basis of this next generation modeling lies in the creation of autonomous agents – human, non-human and systematic, stochastically parameterized to respond and interact with the other agents in their environment on the basis of behavioral rule sets informed by their perceptual capabilities of “seeing,” “hearing,” and otherwise sampling their environment.

An object oriented agent based approach to pedestrian modeling will provide a platform to develop through research deeper understandings of the underlying behaviors that direct human performance and response under stressful situations and can drive facility design and operational protocols to assure competence and capability in untoward situations.

We consider how to address some of the most significant challenges at present in agent-based simulations of mass egress and evacuation. These include more detailed depiction of agents’ movements and decisions; more acurate representations of agents’ behavior, including group movements and the consequent delays; and more attention to long-term effects of injury and exposure, such as respiratory, digestive and immune system disorders and mood disorders, which may have a physical component cause.

In this study, experiments of pedestrian flow through a bottleneck are performed by changing the width of the bottleneck,

b

, from 0.5

m

to 1.4

m

. The time dependence of the pedestrian density in front of and inside the bottleneck is studied. We calculated the velocity of each pedestrian passing through the measurement section inside the bottleneck and the corresponding mean density, and the fundamental diagram is obtained. The relation between the flow rate

J

and the width

b

is studied and compared with the results of other researchers.

Earlier, wayfinding has been studied only in the horizontal context, e.g., in urban navigation. In this paper, we extend route selection preferences to vertical travel arising in multi-storey buildings. We describe our vertical wayfinding model using multi-attribute utility theory. The model is implemented in a simulation tool that is used in designing building transports as well as for evacuation studies in tall buildings. We show how the path selection determines not only the individual decisions but also the end result of the simulation.

The force based models of pedestrian dynamics like the social force model or the centrifugal force model can demonstrate and sometimes explain many features of the collective behaviour of pedestrian crowds as self organization effects based on fairly simple microscopic rules. However, experiments that have been done either gave only collective data like averages of flux and density, or treated only very simple situations like two person interactions. The progress of digital cameras and of image processing during recent years now allows the measurement of pedestrian movements on a ``microscopic'' scale with low costs. Comparing forces derived from measured trajectories with calculated ones for the same situation allows a microscopic analysis of models and shows what parameters are important in special situations. Applying these methods to bottleneck experiments shows the importance of removing the effects of head movement and of stepping from the trajectories before calculating forces.

Pathfinder is an agent-based egress simulator. It uses a combination of steering behaviors and physical constraints to simulate large-scale occupant behavior and evacuation times based on the movement of individual occupants or agents. In this poster, we present algorithms used in Pathfinder, validation and verification results, and a review of Pathfinder’s graphical user interface and output visualization features.

This paper discusses the sonic speed on pedestrian dynamics, in which the flow of pedestrians at bottlenecks shows some similarity to a supersonic airflow in aerodynamics at narrow passage. In order to consider the analogy between these two dynamics, we have investigated the propagation speed of starting wave of pedestrians by performing walking experiments along a line. The propagation speed of successive reaction of pedestrians is regarded as the sonic speed in pedestrian dynamics in terms of propagations of a perturbation. The experimental result shows that the sonic speed in pedestrian dynamics is inversely proportional to the power of the density of pedestrians like in aerodynamics, if the pressure in pedestrian dynamics is assumed as a constant. Moreover, we have found that the power index obtained from the numerical simulations based on the stochastic cellular automaton model is the same with the power index in experiments.

In the context of circulation design for large buildings (e.g. hospitals, airports), the question of sensitivity of the path network against congestions and blockages naturally arises. To date, the answer to this question would require planners to use a simulation package, which is, however, almost never done in the early stages of building design. We therefore propose a novel visual planning tool that enables architects to estimate the impact of disturbances on the building circulation without having to use a simulation package. Our approach is integrated into a common CAD system and visualizes changes in the path-time relationship of adjacent functional areas under the effects of impeded accessibility.

Traditional evacuation induction system is unidirectional and unchangeable, directing to the nearer egress without any change considering the dynamic spread of fire scene. Intelligent evacuation induction system aims at dynamically directing to the safest and most efficient evacuate route far from fire sources, which improves the traditional and passive evacuation idea to active idea. Some intelligent evacuation induction systems have been developed in-and-abroad. However, they can simply realize the function of two-way directional evacuation on the level of local evacuation safety. To achieve real intelligent evacuation on the level of global evacuation safety, a linkage control is realized based on the integration of Human Evacuation Route Optimization Model (HEROM) and the Fire Prevention and Joint Control System (FPJCS). A new intelligent evacuation induction system is established and a case study is illustrated in this poster. It is established on the basis of information distribution between fire spreading scenario and the action status of fire prevention and control facilities in building. Based on dynamic relationship database technology, data transfer is realized between AACA based HEROM and FPJCS. The optimized evacuation route database of building is updated dynamically and timely with consideration of the fire spread scenario. For example, once a fire shutter blocked an evacuation corridor, the ever safe egress became inaccessible, the optimized evacuation route is updated and saved in an order file, the accessibility information of relevant node is recorded and transferred to change the induction direction of the evacuation signs.