Crowd Behavior Research Articles

A dynamic state-based model of crowds

We consider the problem of categorizing, describing and generating the dynamic properties and behaviours of crowds over time. Previous work has tended to focus on a relatively static “typology”-based approach, which does not account for the fact that crowds can change, often quite rapidly. Moreover, the labels attached to crowd behaviours are often subjective and/or value-laden. Here, we present an alternative approach which uses relatively “agnostic” labels. This means that we do not prescribe the behaviour of an individual, but provide a context within which an individual might behave. This naturally describes the time-series evolution of a crowd, and allows for the dynamic handling of an arbitrary number of “sub-crowds”. Apart from its descriptive power (capturing, in a standardised manner, descriptions of known events), our model may also be used generatively to produce plausible patterns of crowd dynamics and as a component of machine learning-based approaches to investigating behaviour and interventions.

Influencing Pedestrian Route Choice Through Environmental Stimuli: A Long-Term Ecological Experiment

Urban centers are getting crowded, public transportation is becoming congested, and mass events are attracting an increasing number of people. Crowd disasters are not rare, and to prevent them the careful planning of pedestrian facilities and collaboration among stakeholders in the organization of events are crucial. When communication and coordination among stakeholders are sufficient, safety can usually be achieved; however, even in such cases, unexpected situations may occur. Automated crowd-control methods are required to address such situations. However, little is known about how crowd behavior can be influenced without direct human intervention. In this study, we investigated the use of environmental stimuli to modify pedestrian behavior (more specifically, route choice) in an educational facility. Colors, lights, signs, and sounds were used to influence route selection. The results show that light and, in part, LED information displays are somehow effective and could be valid candidates to pave the way for automated crowd control systems (especially for night events). The experiment presented here considers low crowd density. However, we believe that this could help encourage the balanced use of space by pedestrians under normal conditions and establish good practices. In turn, this can delay the creation of high densities, which are often the cause of fatalities in crowd disasters, and provide staff with time for intervention.

Adaptive Formation by Pedestrian Small Groups During Egresses

Human crowd behavior is an intriguing example of collective behavior where global dynamics emerge from interindividual interactions. Although most studies of human crowds assume that all individuals are independent, the presence and impact of social groups, such as friends or families, have recently been explored. Spatial structures resulting from group behavior and decision-making processes are key aspects of social groups, and how such structures are formed and the impact of these are largely unknown, especially in the scenario of crowd evacuations. Here we experimentally show that in the vicinity of exits, dyads and triads form a representative formation, which would attenuate a disadvantage of small groups that are typically regarded as moving obstacles toward the evacuation dynamics. In laboratory experiments on crowd evacuations regarding small group behavior, we first observed no noticeable differences of egress time between monads and groups, which is contrary to a naïve intuitiveness but in agreement with recent experimental observations. To resolve this contradiction, a detailed analysis focusing on behavior in front of exits suggests that group members constitute a straight-line structure vertical to the exit. Our findings offer a new perspective on how adaptive small group behavior emerges during crowd evacuation.

A Bi-GRU-DSA-based social network rumor detection approach

Abstract In the rumor detection based on crowd intelligence, the crowd behavior is constructed as a graph model or probability mode. The detection of rumors is achieved through the collaborative utilization of data and knowledge. Aiming at the problems of insufficient feature extraction ability and data redundancy of current rumor detection methods based on deep learning model, a social network rumor detection method based on bidirectional gated recurrent unit (Bi-GRU) and double self-attention (DSA) mechanism is suggested. First, a combination of application program interface and third-party crawler approach is used to obtain microblogging data from publicly available fake microblogging information pages, including both rumor and non-rumor information. Second, Bi-GRU is used to capture the tendency of medium- and long-term dependence of data and is flexible enough to deal with variable length input. Finally, the DSA mechanism is introduced to help reduce the redundant information in the dataset, thereby enhancing the model’s efficacy. The results of the experiments indicate that the proposed method outperforms existing advanced methods by at least 0.114, 0.108, 0.064, and 0.085 in terms of accuracy, precision, recall, and F1-scores, respectively. Therefore, the proposed method can significantly enhance the ability of social network rumor detection.

Human crowd behaviour analysis based on video segmentation and classification using expectation–maximization with deep learning architectures

Human crowd behaviour analysis based on video segmentation and classification using expectation–maximization with deep learning architectures

Macroscopic modeling of social crowds

Social behavior in crowds, such as herding or increased interpersonal spacing, is driven by the psychological states of pedestrians. Current macroscopic crowd models assume that these are static, limiting the ability of models to capture the complex interplay between evolving psychology and collective crowd dynamics that defines a “social crowd”. This paper introduces a novel approach by explicitly incorporating an “activity” variable into the modeling framework, which represents the evolving psychological states of pedestrians and is linked to crowd dynamics. To demonstrate the role of activity, we model pedestrian egress when this variable captures stress and awareness of contagion. In addition, to highlight the importance of dynamic changes in activity, we examine a scenario in which an unexpected incident necessitates alternative exits. These case studies demonstrate that activity plays a pivotal role in shaping crowd behavior. The proposed modeling approach thus opens avenues for more realistic macroscopic crowd descriptions with practical implications for crowd management.

Modeling heterogenous crowd evacuation on stairs in high-rise buildings using a fine discrete floor field cellular automaton model: Accounting for speed and boundary layer variations

The article investigates the impact of crowd heterogeneities on stair evacuation dynamics in high-rise building. Using a fine discrete floor field cellular automaton (FFCA) model, the study explores the often-overlooked influence of crowd variations in mobility and behaviors. The simulation, set in a 21-story staircase, incorporates variations in speeds and boundary layer widths to replicate the heterogenous crowd dynamics. Key findings reveal that the heterogenous group, considering speed variations, exhibit prolonged stair evacuation time, with children and elderly individuals finishing last. Interactions among individuals with different mobilities lead to a 24.1% blocking effect on young adults' evacuation, while young adults enhance the evacuation of children and elderly by 19.0% and 23.8%, respectively. Children on the inner side of stairs lag behind due to the boundary layer, and the entire heterogeneous crowd's evacuation time decreases by 7.0%, highlighting the efficiency boost from boundary layers. A robust linear correlation is identified between specific outflow and the average desired speed for the heterogeneous group, with a consistent density of 0.46ped/m2 near the exit. The proposed model's validity is supported by trajectories consistent with prior observations and empirical data from homogenous crowd evacuation. The study advances our understanding of evacuation behaviors in heterogenous crowds within high-rise buildings during emergencies.

Enhanced Crowd Dynamics Simulation with Deep Learning and Improved Social Force Model

The traditional social force model (SFM) in crowd simulation experiences difficulty coping with the complexity of the crowd, limited by singular physical formulas and parameters. Recent attempts to combine deep learning with these models focus more on simulating specific states of crowds. This paper introduces an advanced deep social force model, influenced by crowd states. It utilizes deep neural networks to accurately fit crowd trajectory features, enhancing behavior simulation capabilities. Geometrical constraints within the model provide control over varied crowd behaviors, adjustable to simulate different crowd types. Before training, we use the SFM to refine behaviors in real trajectories with excessively small distances, aiming to enhance the general applicability of the model. Comparative experiments affirm the effectiveness of the model, showing comparable performance to both classic physical models and modern learning-based hybrid models in pedestrian simulations, with reduced collisions. In addition, the model has a certain ability to simulate crowds with high density and diverse behaviors.

A Coupled SFM-ASCRIBE Model To Investigate the Influence of Emotions and Collective Behavior in Homogeneous and Heterogeneous Crowds

The understanding of crowd behavior dynamics holds immense significance in ensuring public safety across a range of situations, including emergency evacuations and large-scale events. Our research focuses on two primary objectives: investigating the impact of emotions on crowd movement and gaining valuable insights into collective behavior within crowds. To achieve this, we present a coupled model, incorporating an enhanced ASCRIBE model with an agent displacement model. We introduce heterogeneity into our model by incorporating specific mobility laws for different categories of panicked crowds, considering the influence of emotions on both speed and direction. Through numerical simulations, we analyze the model's parameters, observe the behavior of uniform crowds, and explore the collective dynamics within diverse crowds. By conducting comprehensive simulations and analyses, the findings from this study can contribute to the development of more effective crowd management strategies and emergency evacuation protocols.

Research on emergency evacuation behavior rules of pedestrians under fire and explosion accidents in chemical parks considering social relationships

Chemical parks present a high risk of sudden accidents such as fires and explosions, given the significant number of hazardous sources present. This is why emergency evacuation after accidents is so essential. The behavior of crowds during evacuation is complex, and social relationships among pedestrians can have differentiated effects on the evacuation process. However, most of the current crowd evacuation research has focused on life scenarios and natural disaster contexts. Less attention has been paid to the emergency evacuation of crowds affected by social relations in the context of chemical accident disasters. A comprehensive evacuation model that considers the dynamic evolution of crowd transfer both within and outside the factory area is proposed in this paper. Using the Anylogic software based on the social force model, this study constructs emergency evacuation scenarios for crowd in the context of fire and explosion accidents, and toxic gas leakage incidents. Specifically, this study analyzes the impact of three factors, namely inter-personal distance, group size, and proportion of different group sizes in the total population, on the evacuation efficiency. It has been found that when large groups of pedestrians are split into smaller ones for evacuation, the evacuation efficiency can be effectively improved. The study includes calculations of the lethal range of a toxic gas and simulated the process of relocating large crowds of people to nearby shelters. In addition, a strategy for deploying pedestrian groups based on social relationships was developed for the context of mass evacuation. These measures have the potential to optimize the evacuation efficiency of pedestrian groups in chemical parks, including expanding and modifying shelters, diverting dense crowds from residential plots, using interpersonal spacing of 0.1 m, and splitting large groups of pedestrians into smaller ones. The research results can provide a reference for decision making for large-scale group evacuation in chemical parks, and improve the emergency evacuation capability and local risk response capability in chemical parks.

Artificial Intelligence Powered Congestion Free Transportation System Through Extensive Simulations

Intelligent traffic monitoring is a prominent topic of investigation due the emergence of advancements like the Internet interconnected Things and intelligent computers. Combining these technologies will make it easier to methods to aid in making better choices and accelerating urban growth. Intelligent sensing has come to the forefront in recent years due to its capacity to make calculated decisions on its own to address difficult issues. Automatic vehicles and smart gadgets are equipped with sensors that are part of an IoT-based system in order to recognize, gather, and transmit data. Artificial intelligence (AI)-based techniques allow machines to acquire knowledge and keep tabs on their surroundings through continuous sensing. Improvements in variable traffic control strategies for overcrowded cities have numerous positive outcomes, one of which is increased road safety. Since the sensors on which conventional dynamic controllers relied had their own shortcomings, we might use vision sensors (like cameras) to avoid these issues. Image and video-based computing has a lot of potential for measuring traffic volumes. A new traffic management system named Enhanced Transportation Technologies (ETT) is implemented to relieve congestion at the busy intersection after the old one was deemed to be inadequate. The term "intelligent transportation system" (ITS) refers to a group of transportation systems to keep drivers and passengers safe on the road and to facilitate autonomous mobility by optimizing control systems. To further improve urban planning, crowd behavior, and traffic forecasting, dependable AI models have been developed to work in tandem with ITS. Compared to controllers using conventional sensors, the proposed model has been shown through extensive simulations to reduce waiting time and increase movement speed on average.

Review Paper on IoT Thread Detection using Deep CNN Classifier

Abnormal activity will lead to uncommon changes in the crowd behavior. In other words, the crowd motion changes conform to certain rules for valid behaviors, while for abnormal events the motion changes are uncontrolled. The motion-changed rules to detect and localize abnormal behavior in crowd videos. Specifically, we first generate the motion patterns based on the descriptor of collectiveness. Then each frame pair is represented as a transfer matrix whose elements are the difference of a set of motion patterns. Thereafter, the motion-changed rules are constructed in the transformation space using a bag-of-words approach. Finally, the proposed approach measures the similarity between motion-changed rules and the incoming video data to examine whether the actions are anomalous. The approach is tested on the UMN dataset and a challenging dataset of crowd videos taken from the railway station. The experimental results demonstrate the effectiveness of the proposed method for detection abnormal behavior

Crowd Emotion Prediction for Human-Vehicle Interaction Through Modified Transfer Learning and Fuzzy Logic Ranking

In metropolitan environments, unmanned aerial vehicles (UAVs) equipped with video surveillance equipment can monitor crowd behavior and maintain public safety. In high-traffic areas where humans are more likely to make mistakes, a smart city needs modern technology to forecast the behavior of its residents. In order to improve citywide traffic flow, urban transportation systems (UTSs) monitor and learn how people behave in crowds. Using UAVs for video surveillance in smart cities, our research describes a unique way to assess crowd condition, which expands the scope of human-vehicle interactions. Moreover, we use fuzzy logic ranking to improve the system’s ability to detect anomalies in crowds. In order to improve decision-making, a novel deep transfer learning (DTL) technique is applied to the UAV’s received frames. A 98.5% accuracy rate, satisfactory performance, and robustness to population behavior are all characteristics of the proposed integrated model. In UTSs and urban areas, our novel intelligent system analyzes human behavior based on vehicle-human interactions. In areas with low and high traffic congestion, the modified ResNet (mResNet) architecture predicts the crowd’s condition based on fuzzy logic (FLA). Through decision-making based on accurate crowd conditions, best general paths can be selected using the algorithm.

A modified social force model for sudden attack evacuation based on Yerkes–Dodson law and the tendency toward low risk areas

Various evacuation models are the internal support for the movement of pedestrians. Emergencies that occur in public places have the characteristics of suddenness, massive and unpredictability. These characteristics impact pedestrians’ psychological states, ultimately shaping their behaviors. However, the existing models do not fully consider this aspect. Therefore, for sudden attack events, based on the social force model and Yerkes–Dodson law, this paper proposes a modified model which considers the positive and negative effects of psychological state on pedestrians’ behaviors in three dimensions: the realization degree of desired speed, the proximity of attacker to pedestrian, and crowdedness. It also considers the diminishment of repulsion and the attractiveness of low risk areas. A behavioral model based on the exit-reorientation mechanism is proposed to describe the behavior of low risk exit selection according to the relative location of the attacker. The improved model is validated quantitatively and qualitatively, respectively. Simulations based on this model can effectively reproduce the behaviors of crowds in videos. The data of fundamental diagrams from our model fits well with that from videos. We find that the evacuation time is longer when the observation distance within a specific range in the simple scenario. The risk tolerance level of 8 m and 6 m are the critical points affecting the rate of change of survival rate and casualties, respectively. Simulation results indicate that the effect of rational people ratio on evacuation time will vary with risk tolerance level. This study offers insights into the dynamics of crowd evolution during an evacuation following a sudden attack, providing a foundation for exit allocation and emergency crowd guidance.

Real-Time Spatial-Temporal Depth Separable CNN for Multi-Functional Crowd Analysis in Videos

Crowd behavior prediction (CBP) and crowd counting (CC) are the essential functions of vision-based crowd analysis (CA), which play a crucial role in controlling crowd disasters. The CA using different models for the CBP and the CC will increase computational overheads and have synchronization issues. The state-of-the-art approaches utilized deep convolutional architectures to exploit spatial-temporal features to accomplish the objective, but such models suffer from computational complexities during convolution operations. Thus, to sort out the issues as mentioned earlier, this paper develops a single deep model which performs two functionalities of CA: CBP and CC. The proposed model uses multilayers of depth-wise separable CNN (DSCNN) to extract fine-grained spatial-temporal features from the scene. The DSCNN can minimize the number of matrix multiplications during convolution operation compared to traditional CNN. Further, the existing datasets are available to accomplish the single functionality of CA. In contrast, the proposed model needs a dual-tasking CA dataset which should provide the ground-truth labels for CBP and CC. Thus, a dual functionality CA dataset is prepared using a benchmark crowd behavior dataset, i.e. MED. Around 41[Formula: see text]000 frames have been manually annotated to obtain ground-truth crowd count values. This paper also demonstrates an experiment on the proposed multi-functional dataset and outperforms the state-of-the-art methods regarding several performance metrics. In addition, the proposed model processes each test frame at 3.40 milliseconds, and thus is easily applicable in real-time.

Катастрофическое сознание: краткий обзор проблемы

This article presents the relevant for the modern worldview sphere of philosophical reflection on the interaction between consciousness and reality as well as on the way thinking matches its content. The author takes a historical look at the ideas concerning the controversial problems of consciousness, from the views of Confucius and Socrates to the reasoning of modern authors about certain aspects of consciousness leading all subsequent human activity to catastrophic results. In the context of this issue, various approaches to the interaction of human and society with the outside world are demonstrated, as well as the historically developed views on harmonization (the idea of the cosmos as an orderly system of world order in ancient Greece and the laws of Heaven in Chinese pre-philosophy). The paper traces the origins of understanding those aspects of people’s attitude towards property in which one can find the basis for the catastrophic consciousness characteristic of our time. The formation and functioning of “names” in the system of existence is considered, which is presented in the theoretical principles developed by A.F. Losev. In addition, the paper analyses ideas about the possibility of achieving harmony in human relationships in the system of social existence and in the psychology of behaviour of crowds and individuals, as well as the existential views of the latter leading to actions with catastrophic attitudes. Moreover, the theoretical positions of D.S. Sommer and M.K. Mamardashvili’s school are compared. V.P. Vizgin’s article on catastrophic thinking and its main forms of manifestation is considered in detail. This paper uses comparative methods, historical analysis and interdisciplinary synthesis in combination with the techniques of abstraction and analogy

Adoption of Recurrent Innovations: A Large-Scale Case Study on Mobile App Updates

Modern technology innovations feature a successive and even recurrent procedure. Intervals between old and new generations of technology are shrinking, and the Internet and Web services have facilitated the fast adoption of an innovation even before the convergence of its predecessor. While the adoption and diffusion of innovations have been studied for decades, most theories and analyses focus on single and one-time innovations. Meanwhile, limited work has investigated successive innovations while lacking user-level analysis, possibly due to the unavailability of fine-grained adoption behavior data. In this study, we present the first large-scale analysis of the adoption of recurrent innovations in the context of mobile app updates, investigating how millions of users consume various versions of thousands of apps on their mobile devices. Our analysis reveals novel patterns of crowd and individual adoption behaviors, which suggest the need for new categories of adopters to be added on top of the Rogers model of innovation diffusion. We show that standard machine learning models are able to pick up various sources of signals to predict whether users in these different categories will adopt a new version of an app and how soon they will adopt it.

One, Two, Three! Can Everybody Hear Me? Acoustics of Roman Contiones. Case Studies of the Capitoline Hill and the Temple of Bellona in Rome

Abstract Rhetoric was one of the cornerstones of Roman education and public speaking, the essence of being a Roman politician. The speakers attempted to captivate the audience with their style and convince them of their arguments. Studying the audience is therefore just as important as investigating the speakers and their speeches. The aim of this article is to estimate the number of people who could intelligibly hear a speaker delivering a speech from two speaking platforms located in the city of Rome: the podium of the Temple of Bellona in the Campus Martius (in the Late Republican and Late Augustan periods) and the Capitoline Temple. To do this, we built virtual reconstructions of both venues according to the current state of knowledge about them, taking into account the geometry of the space as well as the materials from which they were built. On the models thus prepared, we carried out acoustic simulations for three different levels of background noise (36, 49, and 55 dBA), resulting in Speech Transmission Index maps. The results became the basis upon which we estimated the size of the maximum potential crowds that could hear speech intelligibly, using two methods based on the behaviour of contemporary crowds. We further compared our results with those of previous studies that concern other speaking platforms in Rome.

A Survey on Age and Gender Detection Using Deep Learning

Abstract: The age and gender information are very important for various real world applications such as social understanding, biometrics, identity verification, crowd behavior analysis etc. There are numerous works proposed for age and gender prediction in the past several years. The age and gender detector project aims to develop an advanced AI system capable of accurately determining the age and genders of individuals from images. As we can see, the model outputs are mostly special parts as well as wrinkles, which are important for age and gender prediction. This project utilizes deep learning techniques to analyze special features and patterns, enabling real time identification and classification of age and gender attributes.

Dynamic Disturbance Propagation Model of Pedestrian Panic Behaviors and Lyapunov-Based Crowd Stability Analysis

In public places, pedestrian panic behaviors have received increasing attention due to their dangerous impact on normal pedestrian flow. To address this issue, this study considered crowd panic behaviors as two-dimensional Gaussian disturbances quantitatively triggered by accidents and analyzed the stability of the pedestrian crowd based on Lyapunov criterion. First, this study established a two-dimensional static model for the disturbance pressure in a crowd. Then, a dynamic disturbance–propagation model (DPM) of crowd panic behaviors was proposed based on the conservation law of fluid dynamics. The anisotropy of the disturbance pressure propagation was proven with theoretical derivations and simulation experiments, which kept consistent with ground truth. Further, a stability criterion was proposed for pedestrian crowd flow under disturbances based on Lyapunov theory. To validate the proposed DPM, we simulated a disturbance scenario in the waiting hall of Shanghai Hongqiao Railway Station. Subsequently, the visual disturbance propagation dynamics and crowd state evolution due to a panic behavior disturbance in a pedestrian crowd were investigated; Finally, the experimental results demonstrated that disturbance pressures and pedestrian density fluctuated and diffused with the panic behavior outbreak point as the disturbance center, showing heterogeneous characteristics. This study shows how we can locate the high-risk areas affected by pedestrian panic behaviors in advance, and further help control crowd flow to keep a pedestrian crowd safe in public buildings.