Search

Early crowd dynamics forecasting, such as one week in advance, plays an important role in risk-aware decision-making in urban regions such as congestion mitigation or crowd control for public safety. Although previous approaches have addressed crowd dynamics prediction, they have failed to deal with the scarcity of anomalous events, which results in a large model bias and could not quantify the number of visitors in anomalous crowd gathering. To provide an elaborate early forecast, we focus on the successive properties of importance weighting (IW) to penalize the anomalous data in terms of model bias; however, leveraging the concept of IW is challenging because dividing dataset into normal and abnormal sets is difficult. Motivated by these challenges, we propose CityOutlook, a novel forecasting model based on unbiased regression with importance-based reweighting. To make IW applicable to our approach, we design an anomaly-aware data annotation scheme by utilizing the heterogeneous property of mobility data to determine the data anomaly. We evaluate CityOutlook using the datasets of large-scale mobility and transit search logs. The experimental results show that CityOutlook outperforms the state-of-the-art models on crowd anomaly forecast, providing the same level accuracy in forecasting normal dynamics.

The local crowd density and the crowd distribution estimation tasks are useful but challenging. There are two main problems which limit the performance of existing algorithms. The first one is that there are not enough labeled training samples to build the high-performance estimation model. Another one is that existing methods lack the supports of physical theories of the crowd. To remedy them, a novel crowd density field model is proposed, which is deduced by jointing crowd dynamics theory and social force model. A crowd counting method based on the proposed crowd density field model is introduced to measure the proposed crowd density field model. Extensive experiments confirm the effectiveness of the proposed plan.

With the growing concerns over public safety, the importance of crowd monitoring is being realized by varioussecurity and event management agencies worldwide. Estimation of crowd dynamics can help such agencies in prevention ofany unanticipated accidents or issues. Research on crowd monitoring has been underway since the past few decades. Conventionalcrowd monitoring systems mainly rely on computer vision approach. Due to predominant use of videos/ imagesequences, the existing techniques may raise data privacy concerns. This has led to development of new crowd monitoringtechniques which are privacy preserving and require minimum public participation. This paper aims to serve as a single andsufficient source of information to the concerned researchers on various aspects of crowd monitoring and also provide futuredirections which can be helpful in developing advanced crowd monitoring techniques., info:eu-repo/semantics/published

In this study, pedestrian crowd dynamics at corner turns were investigated by analyzing pedestrian trajectories in a subject experiment for building more reliable, general-purpose, pedestrian simulation models. An experiment under laboratory conditions was conducted wherein a pedestrian crowd walked straight for a short distance before turning into a right-angled corner built with partition walls; the opposite sides were unwalled. Trials were performed with different widths and densities of initial participant positions. Finally, the trajectories of the pedestrians were extracted from a video through computer image analysis. The results demonstrated that pedestrian behavior at corner turns depends on lane position, lane distance (from the wall), and crowd density.

Present simulation and experimental research still have deficiency in depicting the evading behaviour of single pedestrians confronting with an obstacle, which is the basis for the study of crowd dynamics affected by obstacles in real life. Therefore, this study will conduct experiments with a bar-shaped obstacle in the middle of a corridor and explore the corresponding general and particular features of single pedestrians. Particularly, the variation of pedestrian velocity and trajectory under different-sized obstacles will be illustrated. By taking the average velocity and trajectories of the 32 participants, it could be concluded that pedestrians would walk at a velocity of about 1.5 m/s without being affected by the size of obstacle. Besides, pedestrians tend to pass a location about 0.4 meters away from the obstacle edge that is perpendicular to walking direction. Furthermore, pedestrians tend to begin and finish evading the obstacle at locations respectively about 4.40 meters and 4.85 meters away from the obstacle. We also found a heterogeneity in the evading behaviour and pedestrians could be classified into four types accordingly. Results of this study are expected to provide reliable evidence for agent-based modelling in the future.

Event crowd management has been a significant research topic with high social impact. When some big events happen such as an earthquake, typhoon, and national festival, crowd management becomes the first priority for governments (e.g. police) and public service operators (e.g. subway/bus operator) to protect people's safety or maintain the operation of public infrastructures. However, under such event situations, human behavior will become very different from daily routines, which makes prediction of crowd dynamics at big events become highly challenging, especially at a citywide level. Therefore in this study, we aim to extract the deep trend only from the current momentary observations and generate an accurate prediction for the trend in the short future, which is considered to be an effective way to deal with the event situations. Motivated by these, we build an online system called DeepUrbanEvent which can iteratively take citywide crowd dynamics from the current one hour as input and report the prediction results for the next one hour as output. A novel deep learning architecture built with recurrent neural networks is designed to effectively model these highly-complex sequential data in an analogous manner to video prediction tasks. Experimental results demonstrate the superior performance of our proposed methodology to the existing approaches. Lastly, we apply our prototype system to multiple big real-world events and show that it is highly deployable as an online crowd management system.

This paper proposes a critical analysis of the literature addressed to modelling and simulations of human crowds with the aim of selecting the most appropriate scale out of the microscopic (individual based), mesoscopic (kinetic), and macroscopic (hydrodynamical) approaches. The selection is made focusing on possible applications of the model. In particular, model validation and safety problems, where validation consists of studying the ability of models to depict empirical data and observed emerging behaviors. The contents of the paper look forward to computational applications related to the flow crowds on the Jamarat bridge.

This paper proposes an essay concerning the understanding of human behaviours and crisis management of crowds in extreme situations, such as evacuation through complex venues. The first part focuses on the understanding of the main features of the crowd viewed as a living, hence complex system. The main concepts are subsequently addressed, in the second part, to a critical analysis of mathematical models suitable to capture them, as far as it is possible. Then, the third part focuses on the use, toward safety problems, of a model derived by the methods of the mathematical kinetic theory and theoretical tools of evolutionary game theory. It is shown how this model can depict critical situations and how these can be managed with the aim of minimizing the risk of catastrophic events.

International audience; Understanding and predicting the collective behaviour of crowds is essential to improve the efficiency of pedestrian flows in urban areas and minimize the risks of accidents at mass events. We advocate for the development of crowd traffic management systems, whereby observations of crowds can be coupled to fast and reliable models to produce rapid predictions of the crowd movement and eventually help crowd managers choose between tailored optimization strategies. Here, we propose a Bi-directional Macroscopic (BM) model as the core of such a system. Its key input is the fundamental diagram for bi-directional flows, i.e. the relation between the pedestrian fluxes and densities. We design and run a laboratory experiments involving a total of 119 participants walking in opposite directions in a circular corridor and show that the model is able to accurately capture the experimental data in a typical crowd forecasting situation. Finally, we propose a simple segregation strategy for enhancing the traffic efficiency, and use the BM model to determine the conditions under which this strategy would be beneficial. The BM model, therefore, could serve as a building block to develop on the fly prediction of crowd movements and help deploying real-time crowd optimization strategies.

Crowd simulations facilitate the study of how an environment layout impacts the movement and behavior of its inhabitants. However, simulations are computationally expensive, which make them infeasible when used as part of interactive systems (e.g., Computer-Assisted Design software). Machine learning models, such as neural networks (NN), can learn observed behaviors from examples, and can potentially offer a rational prediction of a crowd's behavior efficiently. To this end, we propose a method to predict the aggregate characteristics of crowd dynamics using regression neural networks (NN). We parametrize the environment, the crowd distribution and the steering method to serve as inputs to the NN models, while a number of common performance measures serve as the output. Our preliminary experiments show that our approach can help users evaluate a large number of environments efficiently.

Understanding and predicting the collective behaviour of crowds is essential to improve the efficiency of pedestrian flows in urban areas and minimize the risks of accidents at mass events. We advocate for the development of a & “crowd forecasting system„whereby real-time observations of crowds are coupled to fast and reliable models to produce rapid predictions of the crowd movement and eventually help crowd managers choose between tailored optimization strategies. Here, we propose a Bi-directional Macroscopic (BM) model as the core of such a system. Its key input is the fundamental diagram for bi-directional flows, i.e. the relation between the pedestrian fluxes and densities. We design and run a laboratory experiments involving a total of 119 participants walking in opposite directions in a circular corridor and show that the model is able to accurately capture the experimental data in a typical crowd forecasting situation. Finally, we propose a simple segregation strategy for enhancing the traffic efficiency, and use the BM model to determine the conditions under which this strategy would be beneficial. The BM model, therefore, could serve as a building block to develop on the fly prediction of crowd movements and help deploying real-time crowd optimization strategies.

Crowd model has been very important in the investigation and study of the dynamics of a crowd. Crowd analysis using simulation models such as the Social Force Model (SFM) in representing crowd dynamics is instrumental in this study where experiments with real humans are not practical. However, the SFM lacks features that can represent more detailed individual characteristics and realistic movement when dealing with multi-room environment. Therefore, in this paper, we propose some modifications to the original SFM with a magnetic model approach that also integrates a path finding feature to produce a more realistic simulation of individuals’ movements in the crowd. Results show that common collective crowd behaviours have been produced by the modified crowd model in a multi-room environment.

As the human population grows so does the need to understand human behaviour. One particularly important aspect of human behaviour is how it changes within conglomerations of people i.e. crowds. In this article, a method for modelling crowd behaviour is proposed. This method draws inspiration from the concept of behavioural intention and the related forces of attitudes, influences, and social norms. These topics are first defined and detailed, followed by a survey of related research. Next, the model is presented and adapted to two common crowd dynamic scenarios, which stress a component of behavioural intention. Finally, observations are made about the model, as well as extensions to the models and directions for future research are considered.

Managing the uncertainties that arise in disasters --- such as a ship or building fire --- can be extremely challenging. Previous work has typically focused either on modeling crowd behavior, hazard dynamics, or targeting fully known environments. However, when a disaster strikes, uncertainties about the nature, extent and further development of the hazard is the rule rather than the exception. Additionally, crowds and hazard dynamics are both intertwined and uncertain, making evacuation planning extremely difficult. To address this challenge, we propose a novel spatio-temporal probabilistic model that integrates crowd and hazard dynamics, using ship- and building fire as proof-of-concept scenarios. The model is realized as a dynamic Bayesian network (DBN), supporting distinct kinds of crowd evacuation behavior, being based on studies of physical fire models, crowd psychology models, and corresponding flow models. Simulation results demonstrate that the DBN model allows us to track and forecast the movement of people until they escape, as the hazard develops from time step to time step. Our scheme thus opens up for novel in situ threat mapping and evacuation planning under uncertainty, with applications to emergency response.

The collective behavior of crowds is thought to emerge from local interactions between pedestrians. People widely assumed interaction is heading alignment. The authors tested four dynamical models of alignment against human data from pairs of pedestrians. A “follower” walked with a “leader” while head trajectories were recorded. In the simplest model, the follower's angular acceleration is proportional to the sine of the heading difference (r=0.71); additional damping or delay parameters did not improve the fit. The authors conclude that alignment is controlled by nulling the heading difference. The results provide a cognitively- grounded model of alignment that may be generalized to larger crowds.

The paper describes relation between positions of individuals in a crowd, namely dependence between position of a pedestrian and position of his/her nearest neighbors. Two main characteristics have been analyzed: nth nearest neighbors’ spatial and angular distributions. At first sight, people in human crowd seem to be located randomly, however, the authors findings indicate that there are clearly visible patterns in analyzed characteristics. The authors discover symptoms of strong correlations between position of closely located pedestrians. Simple, local movement rules for pedestrian are proposed to explain observed patterns.

Over the last years multiple variations of the Social Force model have been proposed. While most of the available force-based models are calibrated on observed human movement data, validation for investigating the model characteristics, e.g. variance in parameter values, is still sparse. The authors present a novel methodology for validating Social Force based models which investigates the reproducibility of human movement behavior on the individual trajectory level with real-world movement data. The approach estimates model parameter values and their distribution with non-linear regression on observed trajectory data, where the resulting variances of the parameter values represent the model's validity. The authors demonstrate the approach on a comprehensive (235 pedestrians) and highly accurate (within a few centimeters) set of human movement trajectories obtained from real-world pedestrian traffic with bidirectional flow using an automatic people tracking approach based on Kinect sensors. The authors validate the Social Force model of Helbing and Molnar (1995), Helbing and Johansson (2009) and Rudloff et al. (2011).

Automated object detection and crowd density estimation are popular and important area in visual surveillance research. The last decades witnessed many significant research in this field however, it is still a challenging problem for automatic visual surveillance. The ever increase in research of the field of crowd dynamics and crowd motion necessitates a detailed and updated survey of different techniques and trends in this field. This paper presents a survey on crowd detection and crowd density estimation from moving platform and surveys the different methods employed for this purpose. This review category and delineates several detections and counting estimation methods that have been applied for the examination of scenes from static and moving platforms.

Here, a recent research trend is presented that integrates modeling, simulation, and visual analysis approaches for improving the study of pedestrians and crowd dynamics.

In this work we present a theoretical and experimental study of the natural movement of pedestrians when they walked through a limited and known area. The experimental study was conducted in a shopping center using infrared cameras. We have modeled human pedestrian trajectories by using the principle of least-action principle, which we compared against experimental results.

Video recording right before the Love Parade, Duisburg (Germany) crowd-quake, occurred on 24 July 2010, has been analysed in order to investigate the spatial properties of the crowd (collective case) and those of the single pedestrians in the crowd (individual case). The Grassberger–Procaccia correlation dimension, well known to be able to distinguish patterns in spatial point processes, was used. Our results for this case-study reveal that crowd and single pedestrians are characterized by different spatial behaviour: the whole crowd behaves as a quasi-homogeneous spatial point process through time, with an averaged correlation dimension of about 1.92; while the single pedestrians show a quite large variation of correlation dimensions indicating different spatial patterns, ranging from clustered to quasi-homogeneous.

The modelling and simulation of pedestrians and crowd dynamics require empirical evidences and quantitative data describing the relevant phenomena that models must be able to reproduce. Computer vision can provide several tools both to semi-automatically acquire the demand of a given situation and actually configure a simulation model, as well as to gather information for the sake of model calibration and validation. This paper proposes methods supporting the segmentation and pedestrian counting of crowd flows, the identification and characterisation of main flows in an analysed scene and the detection of social groups in an observed population. The methods are briefly introduced and the achieved results are presented and discussed with reference to the current state of the art.

This paper presents an overview of the mathematical approaches to modeling crowd dynamics by taking into account the complexity of living systems. The contents refer to the classical representation scales (microscopic, kinetic, and macroscopic) and to the mathematical frameworks that can be used for the modeling approach. The analysis focuses on the kinetic scale and, specifically, on developments of the mathematical kinetic theory of particles with heterogeneous behaviors. The existing literature is critically analyzed looking ahead to research perspectives and, more precisely, to a unified modeling strategy in view also of swarm dynamics modeling.

This paper investigates the behaviour of subgroups in crowd dynamics by means of filming and observation. An existing crowd modelling program, CrowdDMX, based on a discrete element model (DEM) has been modified on the basis of observations made in this paper and literature. Each person is represented as three overlapping circles and motion is modelled in a Newtonian manner. It incorporates psychological forces as well as physical forces in a 2D time-stepping environment. The DEM model was modified to include realistic subgroup behaviour, representing people in the crowd desiring to stay together (families, friends, etc.). Subgroup psychological forces were incorporated. The previous model only simulated individuals moving independently, which was unrealistic in some situations as shown by the observation and filming part of the study. The revised program models subgroups realistically including the tendency to avoid subgroup division in cases of contra-flow.

This paper provides a critical review on the experimental studies of complex pedestrian movement behaviors so as to indicate any knowledge gaps in the empirical research on pedestrian crowd dynamics. Types of complex pedestrian movement behaviors are classified, and similarities and dissimilarities between these behaviors are explored. Furthermore, empirical data derived from controlled experiments with human participants are summarized and compared. Also, the use of these data for calibration and validation of pedestrian simulation models are examined. Finally, the differences of pedestrian behaviors between experimental and real-life situations are discussed. Findings from this study are useful for researchers who want to perform experiments or field observations to understand crowd dynamics and for modelers to develop, calibrate and validate their proposed pedestrian simulation models.

Psychological dynamics that are created by the interactions among persons within One-Dimensional Queue, specially the spread of actions from person to person is discussed in this paper. The instability of psychological dynamic factors can be overcome by persons' interactions in a crowd dynamic. Surowiecki's theory states that the individuals in a crowd are not affected by each other in terms of attitudes. This state indicates wisdom of crowd while Gustav Le Bon's theory is in contrast with. The purpose of this paper is to prove that Surowiecki's theory is originated from Gustav Le Bon's theory if there is a large distance between the members of group in the one dimensional queue. This paper proposes a dynamic crowd model by Bergey, Spieser and Davison based on Gustav Le Bon's theory for static crowd model with variable distance interaction; resulting into better control of crowd dynamics using distance interaction.

Unmanned aerial vehicles (UAVs) and unmanned ground vehicles (UGVs) collaboratively play important roles in crowd tracking for applications such as border patrol and crowd surveillance. Dynamic data-driven application systems (DDDAS) paradigm has been developed for these applications to take advantage of real-time monitoring data. In the DDDAS paradigm, one crucial step in crowd surveillance is crowd dynamics modeling, which is based on multi-resolution crowd observation data collected from both UAVs and UGVs. Data collected from UAVs capture global crowd motion but have low resolution while those from UGVs have high resolution information of local crowd motion. This paper proposes an information–aggregation approach for crowd dynamics modeling by incorporating multi-resolution data, where a grid-based method is developed to model crowd motion with UAVs’ low-resolution global perception, and an autoregressive model is employed to model individuals’ motion based on UGVs’ detailed perception. A simulation experiment is provided to illustrate and demonstrate the effectiveness of the proposed approach.

We present a system to conduct laboratory experiments with thousands of pedestrians. Each participant is equipped with an individual marker to enable us to perform precise tracking and identification. We propose a novel rotation invariant marker design which guarantees a minimal Hamming distance between all used codes. This increases the robustness of pedestrian identification. We present an algorithm to detect these markers, and to track them through a camera network. With our system we are able to capture the movement of the participants in great detail, resulting in precise trajectories for thousands of pedestrians. The acquired data is of great interest in the field of pedestrian dynamics. It can also potentially help to improve multi-target tracking approaches, by allowing better insights into the behaviour of crowds.

Fil: Knopoff, Damian Alejandro. Consejo Nacional de Investigaciones Cientificas y Tecnicas. Centro Cientifico Tecnologico Conicet - Cordoba. Centro de Investigacion y Estudios de Matematica. Universidad Nacional de Cordoba. Centro de Investigacion y Estudios de Matematica; Argentina

Recent empirical and theoretical works on collective behaviors based on a topological interaction are beginning to offer some explanations as for the physical reasons behind the selection of a particular number of nearest neighbors locally affecting each individual's dynamics. Recently, flocking starlings have been shown to topologically interact with a very specific number of neighbors, between six to eight, while metric-free interactions were found to govern human crowd dynamics. Here, we use network- and graph-theoretic approaches combined with a dynamical model of locally interacting self-propelled particles to study how the consensus reaching process and its dynamics are influenced by the number k of topological neighbors. Specifically, we prove exactly that, in the absence of noise, consensus is always attained with a speed to consensus strictly increasing with k. The analysis of both speed and time to consensus reveals that, irrespective of the swarm size, a value of k ~ 10 speeds up the rate of convergence to consensus to levels close to the one of the optimal all-to-all interaction signaling. Furthermore, this effect is found to be more pronounced in the presence of environmental noise.

Tragic events in overcrowded situations have highlighted the importance of the availability of good models for pedestrian behaviour under emergency conditions. Crowd models are generally macroscopic or microscopic. In the first case, the crowd is considered to be like a fluid, so that its movement can be described through differential equations. In the second case, the collective behaviour of the crowd is the result of interactions among individual elements of the system. In this paper, we propose a microscopic model of crowd evacuation that incorporates the fuzzy perception and anxiety embedded in human reasoning. A Visual C++ application was developed to evaluate the outcomes of the model. The model was tested in scenarios with presence of a fixed obstacle. Simulation results have been analyzed in terms of door capacity and compared with an experimental study.

Cellular Automata (CA) as bio-inspired parallel computational models of self-reproducing organisms can capture the essential features of systems where global behavior arises from the collective effect of simple components which interact locally. In this aspect, CAs have been considered as a fine candidate to model pedestrian behavior and crowd dynamics in a fine manner. In specific, for crowd modeling, the CA models show evidence of a macroscopic nature with microscopic extensions, i.e. they provide adequate details in the description of human behavior and interaction, whilst they retain the computational cost at low levels. In this paper several CA models for crowd evacuation taking into consideration different modeling principles, like potential fields techniques, obstacle avoidance, follow-the-leader principles, grouping theory, etc. will be presented in an attempt to accomplish efficient crowd evacuation simulation. Moreover, an integrated system based on CAs that operates as an anticipative crowd management tool in cases of medium density crowd evacuation for indoor and outdoor environments is also shown, and its results different real world cases and different environments prove its efficiency. Finally, robot guided evacuation with the help of CAs is also presented. Quite recently, an evacuation system was proposed, based on an accurate CA model capable of assessing the human behavior during emergency situations takes advantage of the simulation output to provide sufficient information to a mobile robotic guide, which in turn guides people towards a less congestive exit at a time.