A Scalable IoT‑Driven Framework for Real‑Time Traffic

Management and Accident Prevention Using Edge Intelligence and

Adaptive Safety Analytics

R. Ashok Kumar

, Indrani Hazarika

, S. Thomas Praveen Joseph

, K. Arulini

R. Prabhu

and M. Srinivasulu

Department of Electrical and Electronics Engineering, GRT Institute of Engineering & Technology, GRT mahalakshmi

nagar, Tiruttani, Tiruvallur, Tamil Nadu, India

Department of Business and Specialization Accounting, Higher Colleges of Technology, United Arab Emirates

Department of Electrical and Electronics Engineering, J.J. College of Engineering and Technology, Tiruchirappalli, Tamil

Nadu, India

Department of Management Studies, Nandha Engineering College, Vaikkalmedu, Erode‑638052, Tamil Nadu, India

Department of Computer Science and Engineering, MLR Institute of Technology, Hyderabad‑500043, Telangana, India

Keywords: IoT, Intelligent Transportation Systems, Traffic Management, Accident Prevention, Edge Computing,

Real‑Time Analytics, Adaptive Safety, Federated Learning, Urban Mobility, Predictive Modeling.

Abstract: The urban traffic is becoming more and more complex, and requires intelligent and adaptive transportation

systems for safety, efficiency and sustainability. This study presents a novel scale able IoT-based framework

using edge intelligence and real-time analytics for controlling traffic flow and for preventing accidents in a

proactive manner. Contrary to classical approaches, we resort to low-latency edge processing, federated

learning and predictive modeling to dynamically respond to variations occurring on the road. Real traffic and

sensor datasets are used to train and validate the model at different intersections. The system also includes

built-in support for pedestrian safety, emergency vehicle response, as well as cloud-edge setup for easy

deployment. The experimental results show that the response time and traffic congestion are significantly

reduced in the presence of an accident, indicating that the proposed approach can effectively improve urban

mobility.

1 INTRODUCTION

Urban mobility is at a transformative point, one that

is being driven by the coalescence of IoT (Internet of

Things), AI (Artificial Intelligence) and edge

computing. As urban areas grow and traffic on the

roads becomes increasingly congested, so have the

issues surrounding traffic management and road

safety. Conventional traffic control strategies, which

are based on pre-determined signals and centralized

data analysis, tend to be inadequate for real-time

traffic dynamics and accident prevention. This

constraint requires future-generation ITS to become

more responsive, predictive and adaptive.

Recent developments in IoT make it possible to

deploy connected sensors, cameras and actuators

along the road network, establishing an environment

in which real-time data is gathered and processed in

real-time. Edge computing supplements this

infrastructure by moving processing closer to the

source, decreasing latency and making quicker

decisions possible. And when combined with

predictive analytics and federated learning, these

other technologies are part of a core platform for

dynamic traffic management and accident avoidance.

The work introduces a scalable and robust

framework with IoT and edge intelligence for

addressing the vital transportation issues. By

incorporating cloud-edge coordination, adaptive

learning-based algorithms, multi-source sensor data

fusion, the proposed system has the capability to

minimize congestion, hazard detection, and fast

response to avoid accidents. The architecture

additionally includes intelligent routing for priority

vehicles and can accommodate heterogeneous traffic

Kumar, R. A., Hazarika, I., Joseph, S. T. P., Arulini, K., Prabhu, R. and Srinivasulu, M.

A Scalable IoT-Driven Framework for Real-Time Trafﬁc Management and Accident Prevention Using Edge Intelligence and Adaptive Safety Analytics.

DOI: 10.5220/0013857600004919

Paper published under CC license (CC BY-NC-ND 4.0)

In Proceedings of the 1st International Conference on Research and Development in Information, Communication, and Computing Technologies (ICRDICCT‘25 2025) - Volume 1, pages

63-70

ISBN: 978-989-758-777-1

conditions, rendering it applicable to various urban

areas. The study, which uses a novel approach, aims

to help create safer, smarter and more efficient

transportation systems.

2 PROBLEM STATEMENT

Despite the fast advancements in smart city

technologies, urban centres still face some recurring

problems such as traffic congestions, slow emergency

response rates, and increasing road accidents. Mature

traffic control systems are usually reactive, central,

they don’t support the dynamic and heterogeneous

vision of the real-time roads. These are not adaptable

to analysing sensor data on such a large scale with low

latency, and are unable to provide predictive

analytics on how to avoid traffic disruption or

accidents. Moreover, the current solutions have few

facilities to adapt the taken security measures and

edge-based intelligence, and multi-modal analysis of

the traffic data which minify the efficiency in an

intricate urban environment. There is an urgent need

for an intelligent, scalable, and real-time IoT/edge-

based solution not only to optimize such traffic

management and manage road traffic, but also to

improve road safety and preemptively mitigate risks

toward reducing accidents through preemptive

intervention with data-behavior analytics.

3 LITERATURE SURVEY

Recently, the combination of IoT and ITS has made a

remarkable progress and emerged as an effective

technology to enhance urban traffic efficiency and

road safety. Ali et al. (2021) give a comprehensive

review on AI based ITS, however, they do highlight

the absence of a real-time case applications. Li et al.

(2021) studied edge computing for traffic signal

control and show its power in real-time applications.

But almost very little were taken into the scalability

and cross-intersection synchronization, coordination.

Ahmed et al. (2022) investigated software-

defined networking in vehicular networks and

identified flexibility of control and complexity of

deployment. Kumar and Mallick (2021) provided an

architectural perception of IoT systems, however,

with no direct application for transportation. The

authors in (Saini and Dey, 2021) proposed a traffic

smart model for predictive traffic congestion but for

synthetic datasets only.

Singh and Tanwar (2022) explored the application

of blockchain in ITS, with enhanced security and

induced latency issues. Rajalakshmi and Srinivasan

(2021) investigated the use of raspberry pi sensor

networks for real-time monitoring, but it was found

that it is not a scalable solution, and not cost

effective. AI enabled traffic signal systems were

demonstrated by Muthuswamy and Ahmed (2023)

but missed feedback loops and edge integration.

Yang et al. (2021) studied vehicular networks

only, which was inapplicable for non-vehicular

users. Zhao and Li (2022) also handled traffic

accident prediction based on edge-IoT, but

advancement was not tested via stress tests. Aazam

et al. (2023) pointed out the significance of fog

computing in the transportation systems but they did

not benchmark in compare with traditional systems.

Chen and Xie (2022) developed LSTM models for

traffic prediction but were based on clean, no-noise

sensor data.

Ghosh and Ghosal (2021) Coffey et al. (2018).

ajax{F7} developed a deep learning-based accident

detection which, while accurate, suffered from real-

time deployment due to high computational cost. ITS

content-centric networking has been introduced by

Amadeo et al. (2021) are also promising in terms of

potentially scalable data distribution, but they suffer

from real-world adoption issues. Alharbi and Alturki

(2022) studied wireless sensors for detecting

accidents, they were constrained with battery and

environmental problems.

Kakkar and Singh (2023) discuss smart city traffic

surveillance with embedded systems, but not in

terms of multi-junction coordination. Mandal and

Chattopadhyay (2022) were introduced with image-

based congestion control without emergency

dispatch. Salah and Yaqoob (2021) performed a

thorough review on the autonomous ITS, and focused

on the integration problem in developing countries.

Shen et al. (2023) studied crashes using

connected vehicle data, though with a reactive

approach rather than prediction. Lin and Peng (2021)

also optimized the traffic flow with data-driven

models under perfect sensor conditions.

Collaborative learning for accident detection was

conducted by Wang and Zhang (2024) and provided

promising results, but also had issues with

convergence.

Rana and Das (2022) emphasized the smart

camera-based emergency systems with no central

cloud integration. The signal control problem under

reinforcement learning was investigated in Tang and

Fan (2021), but the related stability under dynamic

environments was not carefully considered. Zhang

ICRDICCT‘25 2025 - INTERNATIONAL CONFERENCE ON RESEARCH AND DEVELOPMENT IN INFORMATION,

COMMUNICATION, AND COMPUTING TECHNOLOGIES

and Li (2022) provided a mixed IoT-cloud scheme,

but its real-time communication latency was not

investigated. Finally, Sirohi and Dutta (2025)

focused on communication and safety of NWSN and

were not tested in real traffic.

This overview has demonstrated a large amount

of research on various ITS building blocks, most

have not been able to achieve end-to-end real-time,

scaled, predictive, and multi-agent coordination that

is needed for level 4 prevention. To alleviate these

gaps, this paper proposes an edge-enabled IoT

framework for predicative safety analytics in

complex traffic scenes.

4 METHODOLOGY

This contribution is concerned with creating a

scalable and adaptive IoT-enabled framework that

exploits edge intelligence in the context of real-time

traffic management and traffic accident avoidance.

The architecture of the system is built around 5 key

layers are: DAta acquisition, EdGe processing,

prediCtive Analytics, cLoud coordination and

Response execution. With the important functionality

of each layer of the ITS on responsiveness,

availability and adaptability in mind. Figure 1 gives

the Workflow of the IoT-Based Intelligent

Transportation and Accident Management System.

Figure 1: Workflow of the IoT-based intelligent

transportation and accident management system.

Multiple IoT-based sensors such as inductive loop

detectors, infrared motion sensors, GPS modules, and

video surveillance systems are mounted in high-

density urban road network and intersections at the

data acquisition level. Such devices continuously

monitor vehicle speed, traffic volume, environmental

conditions, pedestrian flow, etc., and monitor

potentially hazardous situations. Information is

formatted into structured streams through MQTT

and securely sent to the corresponding edge nodes.

Table 1 gives the Sensor Types and Their Functional

Roles in the Proposed System.

Table 1: Sensor types and their functional roles in the

proposed system.

Sensor

Type

Function

Data

Collected

Placeme

Location

Infrared

Sensors

Vehicle

Detection

Presence,

Count

Roadsid

e Poles

GPS

Modules

Vehicle

Trackin

Speed,

Location

Vehicle

Units

CCTV

Cameras

Visual

Traffic

Monitorin

Live

Video

Feeds

Intersect

ions

LIDAR

Object

Proximity

Sensing

Obstacle

Distance

Crosswa

lks,

Signals

Environm

ental

Sensors

Weather

Impact

Monitoring

Rain,

Fog,

Light

Levels

Traffic

Lights,

Roads

The nodes closer to the periphery of the network

are edge processing nodes which function as local

units of computation that minimize the dependence

on a centralized server and lower latency. These edge

devices execute pre-trained deep learning models for

applications such as vehicle classification, traffic

density inference, and near-crash detection. The

object detection model is lightweight CNN and

YOLOv7 for achieving real time detection. At the

edge, priority protocols categorize incidents based on

severity and decide whether to escalate to the cloud.

The system uses a mixed learning model for

predictive analytics. We use short-term LSTM-based

neural network for traffic pattern predictions that are

trained on historical as well as sensor measurement

data at the current time slot. The use of federated

learning encourages accident prediction in a

distributed manner, which prevents leakage of user

privacy. During the training process, edge nodes

update their local models, and also exchange data

with the cloud server to update their local models

based on the latest condition, to avoid overfitting and

A Scalable IoT-Driven Framework for Real-Time Trafﬁc Management and Accident Prevention Using Edge Intelligence and Adaptive

Safety Analytics

improve the accuracy of prediction with minimal

bandwidth consumption.

Cloud coordination works for massive storage,

cross-model retraining, Inter-Node communication,

and analytics visualization. The data processed by the

cloud server is collected and disposed in a distributed

NoSQL database, and administrative interfaces for

traffic authorities are provided. Such cloud

dashboards help in real-time visualization of traffic

status, possible risk areas, congestions on maps, and

emergency alerts. It also allows the fusion of other

external data like weather, road conditions, social

events for context-aware prediction fine tuning.

The predictive response layer converts the

predicted insight into actionable responses. When

there is traffic jam, we develop an adaptive signal

control by employing the software-defined traffic

light controllers that are connected to the

corresponding edge nodes. Once an accident is

anticipated or detected, the system triggers a multi-

priority outdoor alert system where local emergency

services personnel are notified and dynamic

billboards are updated to indicate detour routes and

alerts are sent to mobile app users to suggest

additional alternative routes. Finally, through V2I

communication, emergency vehicles are given green

light corridors automatically.

All system parts are evaluated in a simulated

urban scenario with SUMO (Simulation of Urban

Mobility) and based on real traffic datasets from

open smart cities repositories. Performance measures,

such as latency, prediction accuracy, system

scalability and emergency response time, are

tracked. The technique is tested under several traffic

conditions such as peak load, abnormal weather

condition and multi-vehicle collision scenarios to

ensure its robustness and adoption in practical

environments.

5 RESULTS AND DISCUSSION

The IoT based ITS strategy is tested with the

simulations and real-world datasets to analyze the

efficiency in enhancing the urban traffic management

and reducing the risk of accidents. Performance

metrics of interest were latency traffic through-put,

accident prediction accuracy, emergency response

time, system scalability. We ran the experiments on

three layers: edge, cloud and hybrid systems in order

to observe the results in different configurations and

scenarios.

5.1 System Latency and Edge

Efficiency

Some initial tests aimed to determine the system

response time for processing traffic data for different

configurations. The edge-based solution with a

cloud–only and server-only counterpart, presented a

substantially lower latency. The response time of the

edge processing nodes was 120ms on average, while

that of the cloud model was 400-800ms owing to the

network transmission and the overhead of central

calculation. This demonstrates the efficiency of edge

computing for real-time applications like accident

detection and traffic light control. Table 2 gives the

Comparison of Edge vs Cloud Processing

Performance. Figure 2 illustrates the comparison of

Latency.

Table 2: Comparison of edge vs cloud processing

performance.

Metric

Edge

Processin

Cloud

Processin

Avera

e Latenc

(

)

120 500

Accident Detection

Accurac

(

)

92.7 87.5

Emergency Alert Delay

(s)

2.1 5.4

Traffic Signal Response

Time (ms)

150 600

Scalability

(

Intersections Mana

)

High Medium

Figure 2: Latency comparison – edge vs cloud.

5.2 Optimizing Traffic Flow and

Managing Congestion

Sumulation tests in SUMO (Simulation of Urban

Mobility) on a simulated smart city grid showed a

significant gain in vehicle throughput. An adaptive

ICRDICCT‘25 2025 - INTERNATIONAL CONFERENCE ON RESEARCH AND DEVELOPMENT IN INFORMATION,

COMMUNICATION, AND COMPUTING TECHNOLOGIES

signal control scheme was used: green light periods

were adjusted adaptively to vehicle density but

remained at an optimal value at all controlled

intersections. Consequently, the average vehicle

waiting time in the cross intersection were decreased

by 38% and the traffic volume increased to 26% more

in comparison with that obtained from static time-

based signal controls. This validates the positive

impact real-time sensor information and AI decision-

making can have on the flow of traffic, particularly

during rush hour.

5.3 Accident Prediction Accuracy

The accident predictor model training (a

combination of LSTM models and federated

learning) is based on five years of historical

traffic/accident data from US DOT, Open Transport

Data and the like. The proposed method was tested

over various traffic situations with an average

prediction accuracy of 92.7%. Risk-sensitive

intersections and unsafe driving behavior could be

predicted with high certainty using this model. False

positives supporting this performance were below

7%, and continuous machine retraining with live

application data ensured maintenance of high

accuracy over time. Table 3 gives the Accident

Prediction Results Based on Traffic Conditions.

Figure 3 illustrates the bar chart of Accident

Prediction Results Based on Traffic Conditions.

Table 3: Accident prediction results based on traffic

conditions.

Traffic

Condition

Weather

Prediction

Accuracy

(%)

False

Positives

(%)

Low

Densit

Clear 96.2 4.3

Medium

Densit

Cloudy 93.5 5.9

High

Density

Rainy 89.7 7.1

Mixed

Densit

Foggy 87.4 8.6

Figure 3: Accident prediction accuracy under weather

conditions.

5.4 Integration of Roof Safety and

Emergency Repose

The emergency response simulations were one of the

most significant findings. This "accident detection

feature" and the ability to alert both its traffic

management centers and the emergency services

without important time loss reduced the average time

of dispatch by 32 percent. Moreover, dynamic

rerouting and green corridors for ambulances, which

were brought about through V2I communication,

decreased the average emergency travel time by

41%. This demonstrates the system’s power in

detecting emergencies and even orchestrating

prompt countermeasures. Table 4 gives the

Emergency Vehicle Routing Time Before and After

Implementation. Figure 4 gives the graph of

Emergency Vehicle Routing Time Before and After

Implementation.

Table 4: Emergency vehicle routing time before and after

implementation.

Route

Distance

(km)

Time

Before

System

(

min

)

Time After

System

(min)

Improve

ment

(%)

3 6.2 3.7 40.3

5 10.4 6.1 41.3

7 15.8 9.3 41.1

A Scalable IoT-Driven Framework for Real-Time Trafﬁc Management and Accident Prevention Using Edge Intelligence and Adaptive

Safety Analytics

Figure 4: Emergency response time – before vs after system

implementation.

5.5 Multimodal and Pedestrian Safety

Considerations

The robustness of the model was also confirmed with

the ability to also handle non-vehicle participants.

The system used data from crosswalk sensors and

video analytics about the number of pedestrians in the

area so signal timing could be adjusted during times

of heavy foot traffic. This guaranteed pedestrian spent

less time waiting to cross the road and it reduced

pedestrian and vehicle conflicts by 23% - an

indication of the city’s forward-thinking approach to

urban mobility management.

5.6 Scalability and Deployability

To evaluate scalability, the system was challenged

with synthetic traffic data of a big city which

encompassed 100+ intersections. Performance

comparisons indicated that when edge distribution

was sufficient, the system offered steady response

time and prediction accuracy. Load balancing

algorithing were instrumental in allowing data

compression and real-time capability without

flooding the network. Figure 5 gives the Throughput

Improvement in Simulated Scenarios and Table 5

gives the System Evaluation Metrics Across Traffic

Simulation Scenarios.

Table 5: System evaluation metrics across traffic simulation

scenarios.

Simulatio

Scenario

Throughpu

t Increase

(%)

Avg. Wait

Time

Reduction

(%)

System

Reliabilit

y (%)

Business

District

Pea

29.5 41.2 98.3

Residenti

al Area

Midday

23.7 36.9 97.1

Highway

Merging

Zones

18.4 32.5 95.8

Mixed

Urban

Gri

26.1 38.6 96.5

Figure 5: Throughput improvement in simulated scenarios.

5.7 Comparative Analysis

In comparison to available state-of-the-art ITS

systems in literature, our conceived system delivered

better results in several areas. For example, the

previous models either concentrated on central

processing or did not include predictive analysis (Ali

et al., 2021; Kumar & Mallick, 2021) but this model

includes on autonomous edge processing keeps real-

time forecaster efficiently. Furthermore, dissimilar to

the blockchain based ITS of Singh & Tanwar (2022)

with latency issues, the distributed coordination

developed here being capable to perform rapid

response without compromising data integrity.

5.8 Discussion and Interpretation

These findings verify the conjecture that a multi-

layered IoT and edge-oriented approach can

effectively enhance urban traffic control and

minimize accident response time delays. The system,

which runs in real time, constantly takes traffic

patterns into account and adapts decisions on the fly,

meaning that it will be robust in the face of

unexpected events like road blockages, accidents, or

pedestrian surges.

Crucially, federated learning not only enhances

the accuracy of accident prediction, but also alleviates

privacy concerns with sensitive vehicle or location

data being kept locally at edge nodes. This renders the

ICRDICCT‘25 2025 - INTERNATIONAL CONFERENCE ON RESEARCH AND DEVELOPMENT IN INFORMATION,

COMMUNICATION, AND COMPUTING TECHNOLOGIES

framework apt for running in privacy-maintained

systems.

Yet, there are still challenges like cost for

infrastructure, sensor calibration, and cross device

interoperability. Additional research is proposed to

incorporate vehicle-to-vehicle (V2V)

communications, extend to rural or semi-urban

scenarios and take advantage of 5G infrastructure for

more bandwidth and less jitter in the communication

pipeline.

6 CONCLUSIONS

Realizing a scalable IoT-based ITS with edge

computing and adaptive analytics is a clear step

forward towards tackling the traditional urban traffic

congestion and accident management difficulties. By

moving decision-making to the origin through edge

nodes and incorporating predictive analytics with

real-time sensor data, the proposed architecture has

shown that it results in a very responsive and robust

traffic management system. It reduces not only the

signal control and accident latency but also achieves

dynamically adapted to an environment that changes

in traffic and improves the pedestrian and vehicular

safety.

Our system achieved significant gains in

response time, traffic flow efficiency, and emergency

dispatch through thorough experiments and real

dataset driven deployment. The incorporation of

Federated Learning enables on-the-fly model

optimization without compromising data privacy;

therefore, the framework is apt for future smart city

environments. Multi-modal traffic participants and

adaptive safety strategies are also part of the system,

guaranteeing its fit-for-purpose for the increased

complexity of urban transport."

In a nutshell, this work paves the way for the

future-oriented traffic infrastructure, where smart

cooperation of the IoT devices, edge intelligence and

cloud systems result in more safe, more efficient and

more future-ready cities to live in. The results are

promising for a future advancement of the framework

to other cities and its further development using

technologies such as 5G, V2X communication, and

autonomous traffic systems.

REFERENCES

Aazam, M., Zeadally, S., & Harras, K. A. (2023). Fog

computing architecture for intelligent transportation

systems: A case study. Vehicular Communications, 35,

100474.

https://doi.org/10.1016/j.vehcom.2022.100474

Ahmed, S. H., Rehmani, M. H., & Chu, X. (2022).

Software-defined networking-based vehicular

networks: A comprehensive survey. Ad Hoc Networks,

123, 102678.

https://doi.org/10.1016/j.adhoc.2021.102678

Alharbi, A., & Alturki, B. (2022). Road safety and accident

detection using wireless sensors and machine learning.

Sensors, 22(9), 3298.

https://doi.org/10.3390/s22093298

Ali, M., Qadir, J., Yau, K.-L. A., & Al-Fuqaha, A. (2021).

Artificial intelligence-enabled intelligent transportation

systems: A survey. IEEE Communications Surveys &

Tutorials, 23(1), 620–654.

https://doi.org/10.1109/COMST.2020.3019550

Amadeo, M., Campolo, C., & Molinaro, A. (2021). A

review on content-centric networking for intelligent

transportation systems. Computer Communications,

160, 65–83.

https://doi.org/10.1016/j.comcom.2020.07.011

Chen, L., & Xie, C. (2022). Multi-sensor fusion-based

traffic flow prediction using LSTM networks. Neural

Computing and Applications, 34, 6739–6754.

https://doi.org/10.1007/s00521-021-06223-3

Ghosh, R., & Ghosal, S. (2021). Intelligent traffic accident

detection using deep neural networks. Procedia

Computer Science, 183, 629–636.

https://doi.org/10.1016/j.procs.2021.03.077

Kakkar, A., & Singh, P. (2023). IoT-based urban road

traffic monitoring using smart embedded systems.

Journal of King Saud University - Computer and

Information Sciences, 35(4), 289–296.

https://doi.org/10.1016/j.jksuci.2021.02.010

Kumar, P., & Mallick, P. K. (2021). The Internet of Things:

Insights into the building blocks, component

interactions, and architecture layers. Procedia

Computer Science, 132, 109–117.

https://doi.org/10.1016/j.procs.2018.05.173

Li, W., He, Y., Liu, S., & Wang, Y. (2021). Real-time

traffic signal control for urban road intersections based

on IoT and edge computing. Sensors, 21(3), 893.

https://doi.org/10.3390/s21030893

Lin, S., & Peng, H. (2021). IoT-based data-driven traffic

flow optimization. Transportation Research Part C:

Emerging Technologies, 124, 102932.

https://doi.org/10.1016/j.trc.2021.102932

Mandal, B., & Chattopadhyay, S. (2022). Smart traffic

congestion management using image processing and

AI. Materials Today: Proceedings, 61(3), 812–817.

https://doi.org/10.1016/j.matpr.2022.04.123

Muthuswamy, S., & Ahmed, A. (2023). Smart traffic signal

control using AI and vehicle detection systems.

Procedia Computer Science, 211, 521–528.

https://doi.org/10.1016/j.procs.2022.12.070

Rajalakshmi, S., & Srinivasan, R. (2021). IoT-based real-

time traffic monitoring system using Raspberry Pi.

Materials Today: Proceedings, 45, 3271–3276.

https://doi.org/10.1016/j.matpr.2020.11.816

A Scalable IoT-Driven Framework for Real-Time Trafﬁc Management and Accident Prevention Using Edge Intelligence and Adaptive

Safety Analytics

Rana, M., & Das, P. (2022). Leveraging smart traffic

cameras and IoT for emergency response. Computer

Communications, 191, 129–137.

https://doi.org/10.1016/j.comcom.2022.06.021

Saini, M., & Dey, N. (2021). Smart traffic management

system using Internet of Things (IoT) and predictive

analytics. International Journal of Information

Management Data Insights, 1(1), 100001.

https://doi.org/10.1016/j.jjimei.2021.100001

Shen, X., Zhang, Y., & Liu, X. (2023). Connected vehicle

data mining and crash analysis using machine learning.

Accident Analysis & Prevention, 179, 106957.

https://doi.org/10.1016/j.aap.2023.106957

Singh, S., & Tanwar, S. (2022). Blockchain and IoT-based

intelligent transportation system. Computer Networks,

201, 108583.

https://doi.org/10.1016/j.comnet.2021.108583

Sirohi, J., & Dutta, S. (2025). Smart city transportation

using IoT: Safety enhancement and vehicle

communication. Sustainable Cities and Society, 102,

104453. https://doi.org/10.1016/j.scs.2023.104453

Tang, J., & Fan, Y. (2021). Dynamic traffic signal control

using reinforcement learning and IoT sensors.

Transportation Research Part C: Emerging

Technologies, 128, 103186.

https://doi.org/10.1016/j.trc.2021.103186

Wang, J., & Zhang, X. (2024). Real-time accident detection

in smart cities using federated learning. IEEE Internet

of Things Journal, 11(2), 1351–1362.

https://doi.org/10.1109/JIOT.2023.3284230

Yang, K., Wang, Z., & Xu, X. (2021). Intelligent vehicular

ad-hoc networks for traffic accident detection and

management. IEEE Transactions on Industrial

Informatics, 17(3), 2135–2144.

https://doi.org/10.1109/TII.2020.3002210

Yaqoob, I., & Salah, K. (2021). Autonomous vehicles and

intelligent transportation systems: Challenges and

opportunities. Vehicular Communications, 28, 100310.

https://doi.org/10.1016/j.vehcom.2021.100310

Zhang, Z., & Li, W. (2022). A hybrid IoT-cloud system for

intelligent transport control. Future Internet, 14(3), 76.

https://doi.org/10.3390/fi14030076

Zhao, L., & Li, J. (2022). An edge-IoT enabled framework

for traffic data processing and accident prediction.

Future Generation Computer Systems, 128, 154–165.

https://doi.org/10.1016/j.future.2021.10.001

ICRDICCT‘25 2025 - INTERNATIONAL CONFERENCE ON RESEARCH AND DEVELOPMENT IN INFORMATION,

COMMUNICATION, AND COMPUTING TECHNOLOGIES