Professional Documents
Culture Documents
A Big Data Framework For Crowd Monitoring in Large Crowded Events
A Big Data Framework For Crowd Monitoring in Large Crowded Events
Article QR
DOI: https://doi.org/10.32350.umt-air.31.04
History: Received: March 20, 2023, Revised: April 26, 2023, Accepted: May 28, 2023, Published:
June 6, 2023
Citation: N. A. Nawaz, M. Abaidullah, and A. Abid, “A big data framework for crowd
monitoring in large crowded events,” UMT Artif. Intell. Rev., vol. 3, no. 1, pp.
51–76, June 2023, doi: https://doi.org/10.32350.umt-air.31.04
A publication of
Department of Information System, Dr. Hasan Murad School of Management
University of Management and Technology, Lahore, Pakistan
Big Data Framework for Crowd Monitoring in Large Crowded
Events
Naeem A. Nawaz1*, Muhammad Abaidullah2, and Adnan Abid3
1
College of Computer and Information Systems, Umm Al-Qura University,
Makkah Al-Mukarmah, Saudi Arabia
2
Conovo Technologies, Lahore, Pakistan
3
Department of Data Science, University of the Punjab, Lahore, Pakistan
ABSTRACT The management of large events with hundreds of thousands of individuals
has remained a challenge over the years. Crushes and stampedes occurring in the events
of mass gathering have swallowed many valuable lives around the world. Considering the
substantial advancement in positional tracking, wearable technology, and wireless
communication, many event organizers are embracing the use of these technologies to get
assistance in managing large events. Intelligent monitoring of crowd movement and
timely analysis of evolving conditions may aid in early detection of critical situations.
The current research aims to propose a big data resource framework to model, simulate,
and visualize the crowd conditions for actual venue settings. A distributed framework has
been presented to monitor the movement and interaction of individuals in large crowded
events through localized sensing and geospatial analysis of massive positional data. The
pilgrimage (Hajj) has been considered as a case study for demonstrating the effectiveness
of the proposed framework. The proposed framework has been with the help of synthetic
data that covered some useful and frequent scenarios based on the case study of
pilgrimage (hajj), which is an annual event involving more than a million people.
INDEX TERMS agent-based modeling, big data, crowd simulation, crowd analytics,
crowd visualization, multi-agent System
I. INTRODUCTION strong chance that people may get
Crowd management in large events with disconnected from their main group.
hundreds of thousands or millions of Particularly, it is a common issue in case
individuals is a critical challenge for of Hajj where most of the pilgrims are old
authorities. For instance, physical and are not accustomed to use cell phones.
management and monitoring of individuals Technology has been used in many
in large events, such as the Olympics, different manifestations to manage these
Hajj, political gatherings, and live concerts large crowds. For instance, computer
is very tedious. The substandard route and vision based techniques have been used to
traffic management in such large events monitor the mobility and behaviors of
may result into congestion which, in turn, crowds. Despite the substantial
creates panic among individuals and advancement in this field, the methods
causes stampedes to occur. Such have turned out to be unproductive when
stampedes have been a major cause of applied at a large scale due to viewing
accidents resulting into injuries and deaths angles and ambiguous detection
of people. Furthermore, in the large scale limitations.
events that span for multiple days, for
instance Hajj and Olympics, there is a
*
Corresponding Author: nanawaz@uqu.edu.sa
UMT Artificial Intelligence Review
52 Volume 3 Issue 1, Spring 2023
Nawaz et al.
On the other hand, the evolving trend of visualization in large events. The
smart or context-aware environments has continuous tracking of individuals at a
led to the proposition of geographically massive scale can be done more efficiently
informative systems. In these systems, through a localized distributed sensing
advanced tracking technologies are used to infrastructure along with the existing
collect quantitative movement data. support of GPS.
Individuals within a crowd are observed as
The current research presents a big data
profiled users. Global Positioning System
resource framework that along with the
(GPS) localization [1] and proximity based
simulation of a physical sensing network
[2] tracking has been used to capture
includes a parallel computing
complex crowd dynamics during an event.
infrastructure to process large sets of
Although, implementing and evaluating
collected positional data for real-time
such systems yet adhere to many
analytics to provide a crowd-monitoring
challenges, such as gathering and
platform for the authorities in any
processing the data from heterogeneous
situation. The proposed framework has
sources and the expense of equipping the
been evaluated by simulating some useful
individuals or places with sensors.
scenarios from a case study of Hajj.
Simulation serves as a means to design
The rest of the study has been organized as
problematic scenarios and the mechanism
follows. Next section discusses the
in order to evaluate the behavior and
relevant literature. Some preliminary
operation of the opted solution or system.
information has been presented in the
Crowd simulation, in particular, attempts
subsequent section which defines the
to give an appropriate abstraction of the
proposed framework. The next section
domain as mentioned above. For the
provides complete details of the proposed
realistic simulation of such events, the
framework and its relevance to one of the
Multi-Agent System (MAS) approach is
frequently organized large-scale event,
considered suitable as the agents are
Hajj. Experiments conducted, for some
expected to move to their goals, interact
useful scenarios, involving spatio-temporal
with their environment, and respond to
analysis of the positional data generated by
each other. MAS is also postulated as a
the simulation for some useful and
preferred approach for emergency
frequently executable scenarios in Hajj,
evacuation [3] because MAS models
have been presented in the next section.
problems in terms of autonomous
While, the last section concludes the
interacting component agents, proving to
research.
be a more natural way of simulating the
un-predictability factor of large crowds. II. RELATED WORK
It is pertinent to mention that crowd Managing and monitoring of large crowded
simulation methods for controlling events has been discussed by many
individual agents in high-density researchers. Some the studies have
crowd,[4] interactive virtual environments, proposed multi-agent systems for crowd
[5] and position-based [6] solutions have simulation and management. While, others
already been proposed. The current study, have worked on managing crowd based on
however, distinguishes itself on the capacity of a venue with the help of
stipulations related to a systematic analysis video data processing and other with the
of the movement of crowds and its help of multi-agent systems. This section
Department of Information Systems
53
Volume 3 Issue 1, Spring 2023
Big Data Framework for Crowd Monitoring…
discusses many different studies which approach (SCA) and split cluster approach
focused on this problem in many different (SpCA) [14].
ways.
An interesting work on managing a massive
A framework for the simulation of crowd according to the capacity of the place
cooperative tasks, performed by agents, was by counting the number of persons getting
proposed by [7]. The proposed solution was in and out from the specific area by using
supportive of four basic actions based on WSN has been presented by [15]. Similarly,
real-life scenarios. Managing workers’ [16] proposed a WSN architecture to find
tasks, avoiding collision, state, and an efficient solution in order to provide
movement of workers was also managed by food items to the massive crowd. The
this framework. They presented an agent- supply of food conducted automatically in
based model to identify the factors which cluster form according to a given limit of
create panic among people. Additionally, time to avoid any food shortage. A real-
they presented a way for evacuating large time based crowd simulation solution
crowds during an unfortunate event [8]. proposed by authors. The presented model
based on the integration of the potential
A data-driven learning technique was
field method and agent-based method.
proposed for representing the information
Different fields are used to calculate the
pertaining to different scenarios of crowd
real-time interactions and prevent collision
simulation. A method based on the low
among agents [17]. Another relevant and
dimensional crowd space technique was
interesting study discussed the self
used to analyze crowd simulation accuracy
evacuation of passengers during panic and
[9].
its solution. For modeling of information,
Some other researchers also developed the spread ripple effect rule is used, a
agent-based simulation techniques by hypothesis is used to model individuals’
extracting the feature of the crowd from a behavior and decision-making during self-
video. For instance, [10] used global path evaluation [18].
planning and collision avoidance strategies
Another mobile phone-based crowd
for evacuating the crowd. A multi agent-
monitoring framework was proposed by
based model presented for evacuating the
[19] by using Wi-Fi and Bluetooth readings
crowd during a violent attack [11]. While,
to estimate the crowd and claim their
[12] proposed a framework based on multi-
solution is the better and low cost. While,
agent reinforcement learning to detect
an automated surveillance system by using
congestion in the crowd. Another agent-
a computational object recognition system
based crowd simulation framework hase
with a video stream was proposed by [20].
been proposed by [13] that employed
This system automatically tracks and
AnyLogic libraries for the simulation of a
identifies the individuals in the crowd by
crowd in runtime environment to simulate
using CNN. Similarly, another framework
emergency evacuation strategies. The
named ‘crowd Probe’ was proposed to
proposed framework based on analytical
monitor the crowd in indoor settings. It
simulation environment. A Cluster
utilizes hidden Markov model for extracting
Verification Model (CVM) proposed by
trajectories of individuals by using Wi-Fi
using a Wireless Sensor Network (WSN) to
monitor [21].
solve the aforementioned problem for
pandemic cases by using single cluster
Another relevant work on the dynamic The proposed model is developed by using
positions and events to manage the large Unity 3D. A video of persons walking and
crowd according to the capacity to avoid interacting is used for simulation and the
causalities was proposed by [22]. They results are presented by using visualization
studied a smart queue approach at various techniques [30]. A framework named
entry points to reduce the waiting time in ‘icrowd’ proposed by authors for
the system as a function of the number of monitoring and tracking the individuals.
multiple entry/exit loops, social distancing, The proposed framework comprised of
and arrival rates of pilgrims [23]. The three layers and a real-time view of
authors designed a WSN based individuals is presented by this framework
identification model by using grouping using real-time location of individuals [31].
techniques and different operational phases In another effort, the authors presented a
to manage the crowd. Optimization of method to link the physics-based
crowd monitoring discussed by authors: a animations and multi-agent modeling to
solution for monitoring of crowd is improve the efficiency of crowd simulation
proposed based on Wi-Fi/Bluetooth [32].
interface. Firstly, a large number of datasets
An improved version of the multi-agent
are prepared and then localization and
reinforcement learning algorithm for
filtration are performed to estimate crowd
learning is done on extracted information
density [24]. In [25], the authors proposed a
from videos [33]. For simulation of crowd,
real-time surveillance-based system namely
an algorithm was proposed to utilize the
smartiSS. The proposed system worked on
trajectories of pedestrians from videos and
the basis of monitoring Mac ids of
simulate these in the form of agents in
individual devices.
simulation [34]. Similarly, a solution for
A spatio-temporal based visualization of simulation of a large number of agents
crowd movement, by using large-scale data based on nonvolatile ram (NVRAM) and
in the form of a transition graph, was agent based architecture for computing the
proposed [26]. A behavior analysis simulation at a large scale was proposed by
approach based on using generative model [35]. An agents based simulation
as Hidden Markov model to help crowd application was proposed by [36]. The
managers in order to make good decisions proposed method comprised of three
in invoking Internet-of-Things (IoT) modules including simulation environment
services. The proposed approach based on for simulating the agents, agent model, and
spatio-temporal flow-blocks for output for visualizing the simulated results.
marginalization of arbitrarily dense flow Mass-motion was used to optimize the
field [27]. The researchers studied about crowd flow rate with density restricted to a
adoption and utilization of fences to guide safe threshold value for efficient crowd
crowd movement. They focused on crowd management. A robust regression model
regarding the optimization of the fence was developed to guide the authorities for
layout for efficient crowd management by the safe and efficient operation of the
using simulation model [28]. The visiting corridor [37]. For the prediction of
researchers designed a smart image stampede in large crowd, authors proposed
enhancement and quality control system for an architecture by using mobile sensing
resource pooling and allocation and IoT network combined with wireless
applied to crowd management systems [29]. multimedia sensing network. They
zones, let’s denote regions as referred in night the place is transformed into a party
(2), (3), and (4). area. Let’s pix denote the point of interest
as pi1, pi2, pi3 …….......pij are different
r1 = z1.r1, z2.r1; ..............z5. r1 (2)
points of interest in the scenario where the
r2 = z1.r2, z2.r2; ..............z9. r2 (3) value of x ranges from 1 to j.
rk = z1.rk., z2.rk...............zs.rk (4) 2) PATHWAYS
It shows that there are multiple zones in a The presence of walls enforces
region. The venue map is used as a constrictions on the area of movement for
background image in simulation calibrated individuals. These walls are created by
with real-word settings to associate pixel applying an alpha image of background in
values with specific places inside the which walls are white where individuals
simulated environment. The positional cannot walk. Whereas, the rest results in
coordinates of moving individuals are that pathways and points of interest to facilitate
of factual surroundings. An event, in the mobility of agents for their movements.
general, can hold many sub-events in Let’s pwk denote the pathways as pw1,
particular. Those sub-events can also be pw2 , pw3 ......... pwy. where the value of k
defined so that a specific place is created ranges between 1 and y. Since, it is a
for the event and when the sub-event ends, known fact that one zone may have many
the place is removed from simulation, pathways, therefore let’s pwi:zj denote the
respectively. The venue of an event pathway i in zone j. Thus, the pathways in
generally, has two substances, a point of different zones would be denoted as
interest for individuals and the available referred in (5), (6), and (7).
path to reach there.
z1 = pw1.z1, pw2. z1…….....pw5. z1 (5)
z2 = pw1.z2, pw2. z2; .......pw9. z2 (6)
zs = pw1.zs, pw2.zs; …......pwy.zs (7)
IV. PROPOSED ARCHITECTURAL
FRAMEWORK
This section explains all the components of
the proposed framework which follows a
layered architecture. The layers and
different components of the proposed
framework have been presented in Figure 2.
FIGURE 1. System Design A distributed positional sensing
1) POINT OF INTEREST infrastructure is placed at a physical layer
which is the quantitative movement data
It is considered as any place inside the source. A big data environment, containing
created world which may be an originating distributed computing platforms and a data
position for some and a destination site for ingestion component, accumulates the
others. Some places remain throughout the processing layer to systematically preserve
simulation, while others are occasional. For and process the anticipated massive geo-
instance, spring festival, in which there are spatial datasets. Multiple visualizations of a
bookstalls in a park during daytime and at moving crowd were proposed at the
recreated to deploy and test the proposed and the list is returned as explained in the
crowd monitoring system’s actual viability algorithm II.
at such a scale. Moreover, possible
scenarios in a large event if simulated and
analyzed beforehand can be useful in
devising the appropriate strategies to
efficiently manage a large crowd on the day
of the event.
A generic context simulator was used [44]
for the basic modeling of entities in a
scenario. The simulator itself is broad in
terms of information sources, however, it is
limited in terms of event-driven activities
and behaviors. The concept was extended
towards immensely crowded events and
activities of individuals that it
comprehends. The simulator works as a
Multi-Agent System (MAS). For the
realistic simulation of such events, MAS
approach is considered suitable as the
agents are expected to move to their goals,
interact with their environment, and
respond to each other. MAS is also
postulated as a preferred approach for
emergency evacuation simulations [3],
because it models the problems in terms of
autonomous interacting component-agents,
which is proving to be a more natural way
of simulating the unpredictability factor of
large crowds. The simulator can generate its
own context and at the same time can also
augment the context coming from real-time
data sensors. The simulated world includes
models for agents, the world, and the
context therein. In order to create random
participants, every participant is assigned a Algorithm I. Create Random Participants
unique ID, group ID, age, nationality, and
activity. The gender of the participant is MongoDB Java driver is selected as Java
assigned colors. Moreover, color is also the development language. It provides both,
assigned to the group leader as explained in synchronous and asynchronous interaction
algorithm I. To create random population with the application. Four shared
world w, the population count is assigned a collections are created to meet the required
variable C, group limit, and group check is functionality of the system. Collections
assigned variable L and G, respectively. with their primary function are shown in
After iteration, random agents are created Table I. Spherical GeoJSON objects are
used to store the locational data. The used to write scripts to analyze the
participant’s position is stored as a point, operational data managed by MongoDB.
while zone bounds are saved as a polygon.
Every participant is assigned a unique ID,
group ID, and role object of MongoDB for
participant role. By using MongoDB query,
every participant with a unique ID, group
ID, role, location, and coordinates of the
participants are inserted in DB as shown in
Figure. 6 to continuously update the
information of individuals by which every
individual can be monitored and located
based on his/her trial. An inclusion
geospatial query, to find points confined in
a polygon, is executed which eventually
returns the number of participants in that
defined region. MongoDB creates a unique
index on ID field by default upon the
creation of a collection. A geo-spatial index
is created on the location field of a record
and a single field index on timestamp.
These indexes collectively increased the
performance of queries. Spark’s Java API is Algorithm II. Create Random Population
TABLE I
SHARED COLLECTION
Collection Primary Function Frequency of Use
To store participant’ data The records are inserted
Participant
record. continuously.
The records are inserted and
Zone To store information on zones. retrieved for zone based
crowd analysis.
To store crowd parametric
The results are stored and
Real-Time Analytics information with respect to
retrieved continuously.
zone or elsewise
TABLE II
OPERATIONAL DATA
Parameters Data- I Data II Data III
Crowd Population 10,000 50,000 100,000
Crowd Observation
180 Min 180 Min 180 Min
Time
Data collection Time
3 Min 3 Min 3 Min
Difference
10000 x 180/3 = 50000 x 180/3 = 100,000 x 180/3 =
No. of Records
600,000 3,000,000 6,000,000
B. TESTING SCENARIOS 3) SCENARIO-III
The proposed framework has been The third scenario is related to data
evaluated on following three frequently analytics. It intends to find people near the
executing scenarios covering different place of any incident at the time of that
aspects of crowd monitoring, management, incident. This is a useful scenario that helps
and analytics. To this end, different in conducting the investigation of an
variations have been incorporated in the incident.
volumes of the data and indexes.
C. EVALUATION
1) SCENARIO-I
The performance of an index is evaluated
The first scenario aims to calculate the on three different volumes of data. The
number of people in a zone at a given point quantum of positional data varies with
in time. This, in turn, can help in calculating respect to the crowd population. The
the total population of a region. performance of a query with or without an
index is measured with respect to its
2) SCENARIO-II
execution time.
The second scenario is a very common
D. OPERATIONAL DATA
problem during the Hajj and helps finding
the lost individuals, as there are a large Queries execution times are given in Table
number of old-age personnel performing III.
Hajj and they get disconnected from their
caretakers very frequently due to many
reasons.
TABLE III
QUERY EXECUTION TIME
Query-I Parameter (Time Milliseconds) Data-I Data-II Data-III
Time (without Index) 550 3093 6484
Time (with Index) 53 788 1230
FIGURE 13. Query-I Participants Present in a Specified Region Within a Certain Time
Frame
Department of Information Systems
69
Volume 3 Issue 1, Spring 2023
Big Data Framework for Crowd Monitoring…
FIGURE 14. Query-I Participants Present in a Specified Region Within a Certain Time
Frame
2) SCENARIO-II LOCATE A LOST and 53 milliseconds with a single field
INDIVIDUAL DURING HAJJ index. Whereas, for Data-II, it took 1596
milliseconds without an index and 23
A compound query, incorporating scenario
milliseconds with a single field index.
II, was executed against different sizes of
While, with Data III, it took 1596
positional data generated by the simulator.
milliseconds without index and 11
It intends to locate the lost individual from
milliseconds with a single field index. It is
his/her current position and through the trial
important to mention that the data model is
of the individual’s movement. Since, it is
designed carefully to accommodate this use
almost impossible to locate a lost individual
case. As a matter of fact, each individual’s
in such a large crowd. The performance of a
trail is stored in a separate file along with
query is measured with respect to its
relevant meta information. Thus, finding an
execution time and the evaluation results
individual’s trail and current location are
are presented in Figure 15 and 16,
possible very quickly with the help of
respectively. The query execution time for
index.
Data-I is 312 milliseconds without an index
FIGURE 15. Query-II Participants Present at a Certain Distance from a Specified Point
FIGURE 16. Query-II Participants Present at a Certain Distance From a Specified Point
3) SCENARIO-III PARTICIPANTS milliseconds without an index and 7800
PRESENT AT A CERTAIN DISTANCE milliseconds with a geospatial index.
FROM THE PLACE OF THE INCIDENT While, for data III it took 52,138
AT THE TIME OF THE EVENT milliseconds without an index and 18,022
milliseconds with a geospatial index point.
A compound query, incorporating scenario
On the other hand, it can be observed from
III, was executed against different sizes of
Figure 18 that the execution time without
positional data generated by the simulator
indexing is approaching to exponential
to locate the individuals and their distance
trend. While, the execution time with index
from specified points, such as boundary
is slightly higher than the linear behavior. It
walls, pathways, entry, and exit points
is pertinent to mention here that this query
defined in the region. This query helps to
is not meant for live data monitoring, rather
monitor the individuals present around the
it will be executed on analytical engine for
place of incident at the time when the event
post-incident investigations. The
took place. It is pertinent to note that such
performance of a query is measured with
queries will be executed on analytical
respect to its execution time and the
engine and not on real-time monitoring site.
evaluation results are presented in Figure 17
The performance of a query is measured
and 18, respectively. Data I includes
with respect to its execution time and the
positional data of 10,000 participants
evaluation results are presented in Figure 17
observed with a time difference of 3 mins
and 18, respectively. Figure 17 shows that
resulting in 0.6 million records. In this case,
for data I, the query execution time was
the query took 312 milliseconds without an
6123 milliseconds without an index,
index, whereas 53 milliseconds with a
whereas 2347 milliseconds with a geo-
single field index to execute.
spatial index. For data II, it took 23,587
[11] Z. Fa, W. Songchun, and S. Zhihua, 2017, pp. 1–5, doi: https://doi.org/
“Evacuation during violent attacks: 10.1145/3018896.3025140
Agent-based modeling and
[17] G. Lu, L. Chen, and W. Luo, “Real-
simulation,” presented at IEEE 9th
time crowd simulation integrating
International Conference on Software
potential fields and agent method,”
Engineering and Service Science,
ACM Transac. Mod. Comput. Simulat.,
Beijing, China, Nov. 23–25, 2018.
vol. 26, no. 4, pp. 1–16, Mar. 2016,
[12] Q. Wang, H. Liu, K. Gao, and L. doi: https://doi.org/10.1145/2885496
Zhang, “Improved multi-agent
[18] L. Hong, J. Gao, and W. Zhu, “Self-
reinforcement learning for path
evacuation modelling and simulation
planning-based crowd simulation,”
of passengers in metro stations,” Saf.
IEEE Access, vol. 7, pp. 73841–73855,
Sci., vol. 110, pp. 127–133, Dec. 2018,
June 2019, doi: https://doi.org/
doi: https://doi.org/10.1016/j.ssci.
10.1109/ACCESS.2019.2920913
2018.05.013
[13] S. Kim, A. Bera, A. Best, R. Chabra,
[19] Y. Yuan, “Crowd monitoring using
and D. Manocha, “Interactive and
mobile phones,” presented at Sixth
adaptive data-driven crowd
International Conference on Intelligent
simulation,” presented at 2016 IEEE
Human-Machine Systems and
Virtual Reality (VR), Greenville, SC,
Cybernetics, Hangzhou, China, Aug.
USA, Mar. 19–23, 2016.
26–27, 2014, doi: https://doi.org/10.
[14] N. A. Nawaz, N. S. Alghamdi, H. 1109/IHMSC.2014.71
Karamti, and M. A. Khan, “An
[20] K. H. Cheong, et al., “Practical
intelligent cluster verification model
automated video analytics for crowd
using wsn to avoid close proximity and
monitoring and counting,” IEEE
control outbreak of pandemic in a
Access, vol. 7, pp. 183252–183261,
massive crowd,” Comput. Model. Eng.
Dec. 2019, doi: https://doi.org/
Sci., 2022, doi: https://doi.org/10.
10.1109/ACCESS.2019.2958255
32604/cmes.2022.020791
[21] H. Hong, G. D. De Silva, and M. C.
[15] N. A. H. Nawaz, H. R. Malik, A. J.
Chan, “Crowdprobe: Non-invasive
Alshaor, and K. Abid, “A simulation
crowd monitoring with wi-fi probe,” in
based proactive approach for smart
Proc. ACM Interact. Mobil Wear
capacity estimation in the context of
Ubiquit Technol., vol. 2, no. 3, pp. 1–
dynamic positions and events,” Adv.
23, Sep. 2018, doi: https://doi.org
Sci. Technol. Eng. Syst. J., vol. 5, no. 6,
/10.1145/3264925
pp. 423–438, Nov. 2020, doi:
https://dx.doi.org/10.25046/aj050651 [22] N. A. Nawaz, A. Waqas, Z. M. Yusof,
A. W. Mahesar, S. and A, Shah, “WSN
[16] N. A. Nawaz, A. Waqas, Z. M. Yusof,
based sensing model for smart crowd
and A. Shah, “A framework for smart
movement with identification: An
estimation of demand-supply for
extended study,” J. Theoret. Appl.
crowdsource management using
Inform. Technol., vol. 95, no. 5, 2017.
WSN,” in Proc. Second Int. Conf. Inter
Things Data Cloud Comput., Mar. [23] N. S. Alghamdi, M. A. Khan, H.
Karamti, and N. A. Nawaz, “Internet of
Things (IoT) enabled smart queuing Comput. Elec. Mathemat. Sci. Eng.
model to support massive safe crowd at Divis., King Abdullah Univ. Sci.
Ka’aba,” Alexand Eng. J., vol. 62, no. Technol, 2023.
12, Dec. 2022, doi: https://doi.org/10.
[30] P. Knob, V. F. de Andrade Araujo, R.
1016/j.aej.2022.06.053
M. Favaretto, and S. R. Musse,
[24] J. Weppner, B. Bischke, and P. “Visualization of interactions in crowd
Lukowicz, “Monitoring crowd simulation and video sequences,” in
condition in public spaces by tracking 17th Braz. Symp. Comput. Games
mobile consumer devices with wifi Digi. Entertain. Oct. 29–Nov. 1, 2018,
inter-face,” in Proc. ACM Int. Joint pp. 250–25, doi: https://doi.org/10.
Conf. Pervasive Ubiquit Comput: 1109/SBGAMES.2018.00037
Adjunct, Sep. 2016, pp. 1363–1371,
[31] S. N. Saeed, A. Abid, E. U. Waraich,
doi: https://doi.org/10.1145/2968219.
S. Atta, A. Naseer, A. A. Sheikh, and
2968414
E. Felemban, “iCrowd—A framework
[25] T. Kulshrestha, D. Saxena, R. Niyogi, for monitoring of identifiable crowd,”
and J. Cao, “Real-time crowd presented at 12th International
monitoring using seamless indoor- Conference on Innovations in
outdoor localization,” IEEE Transac. Information Technology (IIT), Nov.
Mob. Comput., vol. 19, no. 3, pp. 664– 28–30, 2016, pp. 1–7, doi: https://
679, Feb. 2019, doi: https://doi.org doi.org/10.1109/INNOVATIONS.201
/10.1109/tmc.2019.2897561 6.7880036
[26] F. Wu, M. Zhu, Q. Wang, X. Zhao, W. [32] I. Mahmood, M. Haris, and H.
Chen, and R. Maciejewski, “Spatial– Sarjoughian, “Analyzing emergency
temporal visualization of city-wide evacuation strategies for mass
crowd movement,” J. Visualiz., vol. 20, gatherings using crowd simulation and
no. 2, pp. 183–194, 2017, doi: analysis framework: Hajj scenario,” in
https://doi.org/10.1007/s12650-016- Proc. 2017 ACM SIGSIM Conf.
0368-4 Princip. Adv. Discret. Simulat., 2017,
pp. 231–240, doi: https://doi.org/10.
[27] T. H. Noor, “Behavior analysis-based
1145/3064911.3064924
iot services for crowd management,”
Comput. J., June 2022, doi: https://doi. [33] T. Weiss, “Implementing position-
org/10.1093/comjnl/bxac071 based real-time simulation of large
crowds,” presented at IEEE
[28] X.-C. Liao, W.-N. Chen, X.-Q. Guo, J.
International Conference on Artificial
Zhong, and X.-M. Hu, “Crowd
Intelligence and Virtual Reality
management through optimal layout of
(AIVR), San Diego, CA, USA, Dec. 9–
fences: An ant colony approach based
10, 2019, doi: https://doi.org/10.1016
on crowd simulation,” IEEE Transac.
/j.cag.2018.10.008
Intell. Transpor. Syst., 2023, doi:
https://doi.org/10.1109/tits.2023.32723 [34] I. Karamouzas, N. Sohre, R. Narain,
18 and S. J. Guy, “Implicit crowds:
Optimization integrator for robust
[29] A. Hamrouni, “Socially connected
crowd simulation,” ACM Transac.
internet-of-things devices for crowd
Graph., vol. 36, no. 4, pp. 1–13, July
management systems,” M.S. thesis,
Department of Information Systems
75
Volume 3 Issue 1, Spring 2023
Big Data Framework for Crowd Monitoring…