A collective motion model based on two-layer relationship mechanism for bi-direction pedestrian flow simulation

doi:10.1016/j.simpat.2018.03.005

Simulation Modelling Practice and Theory

Volume 84, May 2018, Pages 268-285

https://doi.org/10.1016/j.simpat.2018.03.005 Get rights and content

Highlights

•
A modified social force model is proposed to simulate team-group pattern in bi-direction pedestrian flow.
•
A two-layer relationship mechanism is proposed, leading and social relationships are combined.
•
A method of group vision sharing is proposed to help pedestrians learn other pedestrians around.
•
An aggregate force with visual impact factor is added, and the clustering method is introduced.
•
The collision is reduced by a collective collision avoidance strategy which regards the group as an avoidance unit.

Abstract

In crowd dynamic, relations are existed among some pedestrians, which cause frequent interactions during evacuation, creating collective motion phenomena, such as the most common pattern of team-groups. Besides, collective behavior can make a beneficial effect on the evacuation process. Therefore, this paper proposes a collective motion model to simulate bi-direction pedestrian flow. First, a method of group vision sharing is proposed to help pedestrians learn the crowd around. Based on two-layer relationship mechanism proposed, aggregate force and collective collision avoidance force are added into the original social force formula. The aggregate force is the resultant of two forces, one is the attraction among the leader and team members, and the other one is that among members of groups due to the social relations. Simulation results show that the modified model can reproduce the team-groups collective pattern in real world bi-direction pedestrian flow, and can reduce the collision risk with regarding the group as collision avoidance unit. Furthermore, the evacuation efficiency is improved.

Introduction

The rapid increases in pedestrians and crowded urban infrastructures have made it a significance to explore the pedestrian flow dynamic. Pedestrian traffic congestion brings great inconvenience to people's travels. Moreover, travel delays result in great losses to the social economy and living quality of residents, especially in the public traffic of airplane and high-speed boarding process, a scientific guidance makes a great help to improve the efficiency [1]. In addition, in several large public places, such as the square in an assembly, crosswalk or subway station during rush hours [2], and a crowded concert scene, safe evacuation is a big issue in emergency cases, such as fire and gas leak. The lack of scientific and effective guidance for personnel evacuation results in chaos, which causes stampede and threaten the safety of the lives and properties of pedestrians. Traditional evacuation drills entail enormous manual labor, material resources, and financial resources, and simulating random events and the subconscious interests of pedestrians for their locations is difficult [3]. Using computer simulation technology for scene modeling, path optimization, and crowd movement behavior modeling to simulate the pedestrian flow in public places can contribute to the design of facilities and emergency evacuation guidance [4].

Pedestrian flow is a typical large crowd movement. In pedestrian flows [5], [6], crowd movement includes one and bi-direction, cross, reciprocating, and other forms [7]. The self-organization phenomena of pedestrian flow have distinctive features, such as arching and clogging through the bottle neck and stripe formation of intersecting pedestrians in the bi-direction flow [8], [9], and some scholars came up with methods to observe the occurrence of phase transition [10]. Therefore, the application of evacuation models to simulate pedestrian flow, especially the most typical one—bi-direction pedestrian flow—has been the focus of many scholars.

In the work of bi-direction pedestrian flow, the relationships among pedestrians which are complex and stable in real life should be taken into consideration. Relationships are usually composed of leading and social relationships, which make pedestrians walk together with a herd mentality and is like some biological cluster phenomena somehow [11], [12], [13], [14]. On the basis of the factors above, leading relationship generates team effect and social relations causes small groups effect, these two effects often occur in the process of crowd evacuation. However, most of the existing crowd evacuation simulation models ignore the inter-individual relationships and regard an individual as an isolated particle, which prevent the realistic simulation of crowd evacuation. Thus, a further research on the details of force in team-group effect and in-depth understanding of its behavior and mechanism according to relationship are of vital importance to truly simulating bi-direction pedestrian flow.

This paper proposes a modified social force model (MSFM) that is driven by the two-layer relationship mechanism and a collision avoidance strategy to simulate the self-organized phenomena in bi-direction pedestrian flow. The main contributions of this work are as follows:

(a)
A modified social force model is proposed to simulate the self-organized phenomena of bi-direction pedestrian flow.
(b)
The combination of leading and social relationships forms the two-layer relationship mechanism.
(c)
A method of group vision sharing is proposed to help pedestrians learn other pedestrians around.
(d)
An aggregate force, including visual impact factor is added to the original formula, and the clustering method is introduced.
(e)
The collision in bi-direction pedestrian flow during intersection is reduced by a collective collision avoidance strategy which regards the group as an avoidance unit.

Based on the contributions above, the collective pattern of team-group is reproduced. Each group moves toward the target through the guidance of a leader. Meanwhile, team members gather into small groups according to their social relation. In addition, the lane formation of bi-direction pedestrian flow is reproduced under the collective collision avoidance strategy. A rich set of environmental attributes, specifically a real path environment, is considered. Thus, similar scenes are built to simulate the real scenes. The remainder of this paper is organized as follows. Section 2 mainly introduces related works on pedestrian flow simulation and social force model. Section 3 discusses the original social force model (OSFM). Section 4 presents the MSFM for bi-direction pedestrian flow simulation. Section 5 elaborates the model framework and implementation process. Section 6 shows the efficiency of the proposed model by experimenting on the effects of relationship density and relationship aggregation on the different density flows, and the validity of the model is verified by numerical comparison. The simulation results show that the MSFM can make the pedestrians in a team march with a leader and gather together in small groups to interact with each other, which can achieve orderly and efficient evacuation. The conclusion and future research focus are presented in Section 7.

Section snippets

Related work

In the last decade, numerous simulation models, classified into two categories of microscopic and macroscopic [15], have been presented to elucidate the underlying dynamics of bi-direction pedestrian flow behaviors. Jiang et al. proposed a high-order macroscopic model, which shows that the traffic sonic speed and the group size influence on the width and number of lines [16]. However, macroscopic models cannot commendably describe the local details of pedestrian behavior. Generally, microscopic

Discussion on the original social force model

On the basis of physical mechanics, the forces of the OSFM proposed by Helbing [43] are attributed to the self-driving force generated by the desire of pedestrians to reach their destinations, the forces between pedestrians, and the forces provided by walls in the process of movement. According to Newton's second law, the translation of acceleration and resultant force of pedestrians can be expressed by Formula 1. $m_{i} \frac{d \vec{v_{i}} (t)}{d t} = \vec{f_{i}^{0}} + \sum_{j (\neq i)} \vec{f_{i j}} + \sum_{w} \vec{f_{i w}},$ where $\vec{f_{i}^{0}} = m_{i} \frac{v_{i}^{0} (t) \vec{e_{i}^{0}} - \vec{v_{i}} (t)}{φ_{i}},$ $\vec{f_{i j}} = \vec{f_{i j}^{s}} + f_{i j}$

Two-Layer relationship mechanism

In many existed models especially those which are used to simulate the bi-directional pedestrian flow, pedestrians are considered as isolated and equivalent individuals with no differences in force analysis, which is not consistent with the existence of social relationships in real life. Meanwhile, pedestrians who are more familiar with the surrounding environment can choose a better way and play the important role of a leader in a march.

The video capture presented in Fig. 1 shows that leading

Model framework and implementation process

Based on the above improvements of the SFM, this paper puts forward a MSFM driven by the two-layer relationship mechanism shown in Fig. 8.

The concrete implementation process of the model framework is as follows:

Simulation results and analysis

To verify the effectiveness of the proposed method based on the theoretical model proposed in Section 3, a series of experiments are carried out using VS2012 + OSG as the experimental platform of the crowd evacuation simulation system. In this chapter, the modified model is tested and verified in contrast with other literature models and the real bi-direction pedestrian flow. Section 6.1 illustrates the contrast test, and Section 6.2 shows the simulation and real experiment results. The width

Conclusions

To solve the problem of emergency evacuation and to reproduce the self-organization phenomenon of the bi-direction pedestrian flow, this paper combines leadership and social relationships to propose the two-layer relationship mechanism. Based on the force formula of the OSFM, aggregate force is added with the method of group vision sharing and relations. Thus, the pedestrians with close relationships can find each other and gather in small groups under the leadership of the leader, and more

Acknowledgment

This research is supported by the National Natural Science Foundation of China (61472232, 61272094 and 61572299) and by the project of Taishan scholarship.

References (64)

T.Q. Tang et al.
An aircraft boarding model accounting for passengers’ individual properties
Transp. Res. Part C
(2012)
X. Chen et al.
A multiagent-based model for pedestrian simulation in subway stations
Simul. Modell. Pract. Theory
(2017)
A. Sagun et al.
Computer simulations vs. building guidance to enhance evacuation performance of buildings during emergency events
Simul. Modell. Pract. Theory
(2011)
L. Fu et al.
Multi-grid simulation of counter flow pedestrian dynamics with emotion propagation
Simul. Modell. Pract. Theory
(2016)
Y. Jiang et al.
Macroscopic pedestrian flow model with degrading spatial information
J. Comput. Sci.
(2015)
T. Vicsek et al.
Collective motion
Phys. Rep.
(2012)
X. Zheng et al.
Modeling crowd evacuation of a building based on seven methodological approaches
Build. Environ.
(2009)
Y.Q. Jiang et al.
A higher-order macroscopic model for bi-direction pedestrian flow
Physica A
(2015)
Y. Han et al.
Modified social force model based on information transmission toward crowd evacuation simulation
Physica A
(2017)
M. Muramatsu et al.
Jamming transition in pedestrian counter flow
Physica A
(1999)

R.Y. Guo et al.

A mobile lattice gas model for simulating pedestrian evacuation

Physica A

(2008)

X. Guo et al.

A heterogeneous lattice gas model for simulating pedestrian evacuation

Physica A

(2012)

H. Xi et al.

Two-level modeling framework for pedestrian route choice and walking behaviors

Simul. Modell. Pract. Theory

(2012)

T.Q. Tang et al.

Modeling pedestrian movement at the hall of high-speed railway station during the check-in process

Physica A

(2017)

V.J. Blue et al.

Cellular automata microsimulation for modeling bi-directional pedestrian walkways

Transp. Res. Part B

(2001)

R.Y. Guo et al.

A simulation model for pedestrian flow through walkways with corners

Simul. Modell. Pract. Theory

(2012)

Z. Wang et al.

Team-moving effect in bi-direction pedestrian flow

Physica A

(2012)

J. Wei et al.

Experiment of bi-direction pedestrian flow with three-dimensional cellular automata

Phys. Lett. A

(2015)

T.Q. Tang et al.

An evacuation model accounting for elementary students' individual properties

Physica A

(2015)

C. Burstedde et al.

Simulation of pedestrian dynamics using a two-dimensional cellular automaton

Physica A

(2001)

M. Isobe et al.

Experiment and simulation of pedestrian counter flow

Physica A

(2004)

Y. Tajima et al.

Pattern formation and jamming transition in pedestrian counter flow

Physica A

(2002)

M. Muramatsu et al.

Jamming transition in pedestrian counter flow

Physica A

(1999)

W.G. Weng et al.

A behavior-based model for pedestrian counter flow

Physica A

(2007)

J. Liang et al.

An extended small-grid lattice gas model for pedestrian counter flow

Procedia Eng

(2013)

Y. Li et al.

A grouping method based on grid density and relationship for crowd evacuation simulation

Physica A

(2017)

T. Kretz

On oscillations in the social force model

Physica A

(2015)

W. Li et al.

The trace model: A model for simulation of the tracing process during evacuations in complex route environments

Simul. Modell. Pract. Theory

(2016)

P. Ma et al.

The escape of pedestrians with view radius

Physica A

(2013)

D.R. Parisi et al.

A modification of the social force model can reproduce experimental data of pedestrian flows in normal conditions

Physica A

(2009)

H. Liu et al.

Crowd evacuation simulation approach based on navigation knowledge and two-layer control mechanism

Inf. Sci.

(2018)

M. Nasir et al.

A genetic fuzzy system to model pedestrian walking path in a built environment

Simul. Modell. Pract. Theory

(2014)

Cited by (27)

A heterogeneous Social Force Model for Personal Mobility Vehicles on futuristic sidewalks
2024, Simulation Modelling Practice and Theory
Electric scooters are becoming popular in public spaces, and autonomous robots will join soon. However, integrating these Personal Mobility Vehicles (PMV) without proper provisions challenges the safety and comfort of all users. While Social Force Model (SFM) commonly replicates pedestrians’ movements, directly applying it to PMVs is challenging and inaccurate. We propose a heterogeneous SFM considering the dynamic personal spaces of various agents on futuristic sidewalks, addressing the impracticalities of SFM. Additionally, subjective safety estimation relaxes the constant desired-velocity assumption, and the influence weight reduces the complexity by omitting pairwise calibration. Experiments calibrate the model for e-scooters, validating it in realistic scenarios with multiple e-scooters passing through pedestrians. The proposed model has higher accuracy than previous models regarding behavioral naturalness metrics. In addition, the models’ performance in replicating experimental observations is analyzed. This research contributes to safer and more efficient transportation with PMVs, particularly e-scooters, and provides a novel approach to modeling multi-type agents on heterogeneous sidewalks.
Airport passenger building assessment of circulation facilities for smooth traffic and efficient design
2022, Simulation Modelling Practice and Theory
The privatisation of airports and the development of commercial activities inside passenger terminal buildings raise questions about how to best manage floorspace to generate revenue, while preserving adequate circulation capacity. Regarding privatised airports, the regulators are challenged to assess the proposed designs in terms of capacity and level of service, most of the time relying on peak hour-based approximations. This study aims to deliver a generic framework for the assessment of airport passenger terminal circulation facilities, providing baseline results that can be used for planning purposes. The Four-Step Model for the Airport Terminal (FSMAT) methodology herein proposed is inspired by the classic Four-Step Model (FSM) and focuses on the horizontal and vertical circulation facilities. The FSMAT model relies on three phases: (i) peak traffic estimation; (ii) mode choice; and (iii) link capacity assessment, featuring an integrated and flexible approach. The FSMAT phase i, especially, was systematically evaluated by means of a queueing network simulation with data from US and Brazilian airports. The results indicate that circulation facilities cannot be designed from hourly movements, as passenger flows are subject to peak fluctuations. Finally, a set of regression models is developed for the prompt estimation of peak passenger traffic, shedding light on the understanding of its influencing factors.
A model for flexible representation of social groups in crowd simulation
2021, Computers and Graphics (Pergamon)
Citation Excerpt :
We can also observe that, because there is no representation of collisions in the model of Moussaid et al. [5], the agents interpenetrate one another many times, which is an undesirable artifact. A similar experiment in a bidirectional corridor was done by Qin et al. [19] (Fig. 9). In that work, although they model groups with leaders that make such groups move orderly under a leadership in relatively closely linked teams, the absence of a leadership hierarchy prevents the groups from having a more organized structure such as the ones shown in Fig. 8d.
Taking into consideration the behavior of social groups, such as groups of friends, couples and families, is very important in crowd simulation. Existing approaches of group modeling offer little flexibility for customizing the structure of the groups and the relationships between their members, which limits their representation power. Often, those approaches also have limitations regarding their ability to represent groups of arbitrary sizes, as well as to represent certain aspects of the groups’ cohesion, such as the ability of modeling how their members will walk together and whether they will follow the same routes. This work proposes a way of modeling groups with different structures, in which roles of leadership can be assigned to some of their members. The approach can also represent groups of arbitrary sizes that may consist of subgroups, and treat different levels of relationships between their members. Moreover, the groups can be modeled in such a way that their members show either a strong cohesion with one another or a loose cohesion, in which the members behave more autonomously when dealing with collision avoidance and choice of routes. Nevertheless, the sense of group is always present as the model offers the ability of regrouping whenever one or more members of a group are left behind. The model also includes a new personality factor that represents the urge of the group members to move towards their destinations when they experience a strong impedance from the crowd. The approach is based on states that the pedestrians can assume based on their profiles and on the social forces model.
Image smoothing based on histogram equalized content-aware patches and direction-constrained sparse gradients
2021, Signal Processing
For better removing textures with larger gradients while preserving structural edges with smaller gradients, we propose an image smoothing method based on histogram equalized content-aware patches and direction-constrained sparse gradients. In order to better smooth the boundary concentration regions while maintaining the continuity of boundary pixels, a content-aware patching technology with boundary constraints is proposed. The irregular patches are represented by the smallest rectangular bounding boxes to reduce the computational complexity. Based on edge information, patches are divided into edge-patches and non-edge-patches. Histogram equalization is used to improve the edge gradient of patches with structural edge concentration, while the image decomposition is used to reduce the gradient of texture details. Taking the edge information as the smoothing factor, each patch is smoothed via direction-constrained sparse gradients. The entire image needs to be further smoothed to remove residual texture details. Experimental results show that new method has better visual effects in retaining structural edges and removing texture details, and has many applications, including edge detection, image abstraction, detail enhancement, and texture mapping.
A modified social force model for high-density through bicycle flow at mixed-traffic intersections
2021, Simulation Modelling Practice and Theory
Citation Excerpt :
Much research effort has been put on expanding the set of forces through adding specific collision-avoidance/behavioral forces or changing the parameters based on their simulation implementation [32-36]. This is often beneficial for formulating traffic users’ dynamics, therefore, these modified models can successfully capture a large amount of collective phenomena involving pedestrians, such as lane formation [37] and emergency evacuation [14,38-40]. However, complex situations with the behavior process exceed the capability of the force-based model [21].
Mixed-traffic intersections with high-density bicycle traffic flow are one of the most common bottlenecks in the urban traffic network. Due to the shared space and frequent interactions among heterogeneous road users, through bicycle flow shows an obvious tendency of lateral dispersion, resulting in a reduction in traffic efficiency and safety at intersections. A precise evaluation of its impacts on intersection efficiency and safety is necessary, while the microscopic simulation model is a useful tool to fulfill the need. However, existing microscopic simulation models of through bicycles simplify cyclists’ behaviors and interactions with other road users, and cannot reproduce the lateral dispersion with dynamic variations accurately. To address the shortcomings, this paper proposes a modified social force model to simulate the through bicycle flow at mixed-traffic intersections. A dynamic boundary model and the behavior force model embedded with the decision process are introduced. The dynamic boundary model captures the dynamic characteristics of the lateral dispersion for different cycles under various traffic context. The behavior force model composes four interactive behaviors, i.e. freely moving, following, overtaking, and merging behavior, and rule-based behavior decision models are integrated with force-based approaches to represent interactions and bicycle dynamics. The proposed model has been calibrated based on 743 through bicycle trajectories at a signalized intersection in Shanghai, China. The simulation results indicate that the proposed model is more capable of reproducing the realistic motion features of through bicycles, and show higher simulation accuracy than the original social force model.
An extended dynamic model for pedestrian traffic considering individual preference
2021, Simulation Modelling Practice and Theory
Citation Excerpt :
And the third stream uses an optimization-based methodology to develop models which determine optimal evacuation plans or design solutions [1]. Different modelling and solution techniques are proposed for evacuation problems and design of pedestrian facilities such as continuum models, network-based models, cellular automata models, agent-based models, social-force models, game-theoretic models, and so on [3–18]. Among these models, the cell transmission model (CTM) by Daganzo [19,20] is seen as a bridge of macroscopic and microscopic models [21–23].
Complexity, dynamics and mutability are basic characteristics of pedestrian evacuation. However, most models for pedestrian traffic still lack an explanatory mechanism of the individual choice process. This study concerns a cell transmission model (CTM) for pedestrian traffic. But we focus on a new issue, namely, how to illustrate the decision-making process of each individual pedestrian. The decision field theory (DFT), a cognitive model to capture the deliberation process in decision making, is introduced to the cell transmission model and an extended dynamic pedestrian flow assignment model based on DFT and CTM is proposed. Pedestrian attitudes towards distance and crowdedness are considered and the individual preferences of pedestrians in the dynamic decision-making for the evacuation process are explored. Numerical studies show that the initial pedestrian distributions can have an effect on the evacuation process and different individual preferences of pedestrians will lead to different pedestrian choices. Even in the same cell, pedestrians’ route and exit choices may be different. Furthermore, some interesting phenomena can be observed, which illustrate that the preference reversals can be depicted in our model.

View all citing articles on Scopus

View full text

A collective motion model based on two-layer relationship mechanism for bi-direction pedestrian flow simulation

Highlights

Abstract

Introduction

Section snippets

Related work

Discussion on the original social force model

Two-Layer relationship mechanism

Model framework and implementation process

Simulation results and analysis

Conclusions

Acknowledgment

Transp. Res. Part C

Simul. Modell. Pract. Theory

Simul. Modell. Pract. Theory

Simul. Modell. Pract. Theory

J. Comput. Sci.

Phys. Rep.

Build. Environ.

Physica A

Physica A

Physica A

Physica A

Physica A

Simul. Modell. Pract. Theory

Physica A

Transp. Res. Part B

Simul. Modell. Pract. Theory

Physica A

Phys. Lett. A

Physica A

Physica A

Physica A

Physica A

Physica A

Physica A

Procedia Eng

Physica A

Physica A

Simul. Modell. Pract. Theory

Physica A

Physica A

Inf. Sci.

Simul. Modell. Pract. Theory