Fastag Fraud Detection System

Sakshi Dodawad

, Sachin Somappa Sidnal

, Shraddha G Shahapurkar

, Shreya Kadakol

Shweta Madiwalar

and Neelam Somannavar

Department of Electronics and Communication Engineering, KLE Technological University, Dr. MSSCET Campus,

Belagavi, Karnataka, India

Keywords:

FASTag, Fraud Detection, Machine Learning, Python, Anomaly Detection, Electronic Toll Collection,

Transaction Security.

Abstract:

The increasing adoption of FASTag for electronic toll collection has streamlined vehicular payments across

toll plazas in India but has also introduced new risks of fraud. This paper presents a machine learning-based

approach for detecting fraudulent transactions in FASTag systems, implemented in Python. By analyzing

transaction patterns, identifying anomalies, and employing classiﬁcation and anomaly detection algorithms,

our proposed system detects potential fraud in real time. This solution aims to reinforce the security and

integrity of the FASTag ecosystem, safeguarding against unauthorized usage and ﬁnancial loss. Our study

includes a review of existing fraud detection methods in digital payment systems, followed by an evaluation of

our approach through performance metrics such as accuracy and precision. Experimental results demonstrate

the system’s effectiveness in identifying suspicious activities, thus providing a valuable tool for enhancing

security in electronic tolling infrastructure.

1 INTRODUCTION

The FASTag system, introduced by the National

Highway Authority of India (NHAI), has revolution-

ized toll collection by enabling RFID-based, cashless

electronic toll collection (ETC) on national highways.

This system, which mandates the use of FASTag for

all vehicles, has signiﬁcantly improved trafﬁc ﬂow

by reducing congestion at toll plazas and has pro-

moted transparency and accountability in tolling op-

erations. With the widespread adoption of FASTag,

millions of transactions are processed daily across

India’s extensive highway network, making toll col-

lection faster, more efﬁcient, and more convenient.

However, the rapid expansion of the FASTag system

has led to new security challenges, including the risk

of fraudulent activities that undermine the trustwor-

thiness and integrity of the ETC framework. These

frauds include unauthorized account access, cloning

of FASTag accounts, and exploitation of system vul-

https://orcid.org/0009-0005-1065-4774

https://orcid.org/0009-0005-4281-5579

https://orcid.org/0000-0001-7110-8906

https://orcid.org/0009-0001-0016-625X

https://orcid.org/0009-0002-7428-3739

nerabilities, posing a risk to both toll operators and

users .The rise in FASTag-related fraud calls for ef-

fective and efﬁcient fraud detection mechanisms to

identify and mitigate risks in real-time. Traditional

rule-based approaches often fall short when dealing

with high-volume, dynamic transaction data due to

the complex and evolving nature of fraudulent pat-

terns. To address this gap, this paper presents a ma-

chine learning-based framework for fraud detection in

FASTag transactions, utilizing techniques such as lin-

ear regression and logistic regression. Machine learn-

ing (ML) offers powerful tools to detect anomalies

within high-frequency transaction data, with super-

vised models like linear and logistic regression prov-

ing effective in binary classiﬁcation tasks and predict-

ing probabilities based on transaction patterns. By

analysing features derived from transaction histories,

our system seeks to distinguish between legitimate

and suspicious transactions, improving the overall se-

curity of the FASTag ecosystem .The proposed so-

lution centres on developing a comprehensive fraud

detection model using Python-based machine learn-

ing techniques tailored for real-time FASTag trans-

actions. In the preprocessing stage, linear regression

aids in understanding and normalizing transactional

data patterns, enhancing feature engineering by iden-

904

Dodawad, S., Sidnal, S. S., Shahapurkar, S. G., Kadakol, S., Madiwalar, S. and Somannavar, N.

Fastag Fraud Detection System.

DOI: 10.5220/0013607000004664

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

In Proceedings of the 3rd International Conference on Futuristic Technology (INCOFT 2025) - Volume 2, pages 904-909

ISBN: 978-989-758-763-4

tifying underlying trends and detecting anomalies in

transaction amounts and frequencies. This analysis

helps to capture seasonal variations and scale data ef-

fectively, ensuring that key patterns associated with

fraud are highlighted. Logistic regression, renowned

for its binary classiﬁcation capabilities, is employed

as the primary model to distinguish between “fraud-

ulent” and “non-fraudulent” transactions. By train-

ing the model on historical labelled data, it gains the

ability to recognize prevalent fraud tactics, such as

account cloning, unauthorized transactions, and un-

expected frequency of usage, thereby improving its

detection accuracy. To further enhance accuracy, the

model undergoes iterative training and hyperparam-

eter optimization, ﬁne-tuning aspects like regulariza-

tion and feature weights to reduce false positives and

increase detection reliability. Additionally, the model

continuously learns from new data, enabling it to dy-

namically adapt to evolving fraud patterns. Such an

approach not only enhances the operational security

of the FASTag system but also aligns with regulatory

compliance requirements, strengthening the overall

trust and robustness of the electronic toll collection

ecosystem. The remainder of this paper is structured

as follows: Section II provides a literature review on

existing fraud detection techniques in electronic pay-

ment and toll collection systems. Section III outlines

the methodology, describing data preprocessing, fea-

ture extraction, and the machine learning algorithms

employed. Section IV presents experimental results,

including model performance and accuracy metrics,

and discusses the implications of these ﬁndings. Fi-

nally, Section V concludes the paper and suggests po-

tential future directions for enhancing this fraud de-

tection system. Through this work, we aim to con-

tribute to the growing body of literature on the ap-

plication of machine learning in securing digital pay-

ment and tolling infrastructures, ultimately promoting

a safer, more reliable FASTag ecosystem

2 RELATED WORKS

1. (Oza, ) Enhancing Integrity of Toll Gates:

Fastag Fraud Detection: delves into the application

of blockchain technology to strengthen data shar-

ing, security, and privacy within Intelligent Trans-

portation Systems (ITS), with a focus on toll collec-

tion and FASTag systems. Recognizing vulnerabili-

ties in existing tolling methods, the authors propose

a blockchain-based framework to counteract fraud

and enhance data integrity by decentralizing data

management and ensuring transparency through im-

mutable transaction records. The paper highlights

how blockchain can streamline economic efﬁciency

within ITS by reducing the need for third-party ver-

iﬁcations, lowering operational costs, and accelerat-

ing transaction processes. Additionally, integrating

blockchain with the Internet of Vehicles (IoV) is ex-

plored as a way to create a cohesive and secure ITS

ecosystem. This integration would enable real-time

data exchange between vehicles, toll stations, and

other ITS components, thereby enhancing data pri-

vacy and transparency in tolling operations. By po-

sitioning blockchain as a foundational layer, the pa-

per suggests a future in which tolling systems are not

only fraud-resistant but also more secure, resilient,

and scalable.

2. (Bhavar et al., 2023) Fastag Fraud Detec-

tion: A Literature Survey: provides a comprehen-

sive overview of existing research and methodolo-

gies applied to detect fraudulent activities within the

FASTag electronic toll collection system. This liter-

ature survey examines various types of fraud asso-

ciated with FASTag transactions, such as unautho-

rized access, account cloning, and system exploita-

tion. The authors analyse both conventional and ma-

chine learning-based approaches that have been pro-

posed for fraud detection in digital payment systems,

highlighting their strengths, limitations, and appli-

cability to FASTag. The survey also explores var-

ious machine learning techniques, including super-

vised and unsupervised learning methods, that have

been effectively applied to similar fraud detection sys-

tems, such as credit card and online payment fraud.

By identifying gaps in the current research, the au-

thors emphasize the need for robust, real-time fraud

detection frameworks that can adapt to evolving fraud

tactics. The paper concludes by suggesting poten-

tial directions for future research, including the inte-

gration of advanced machine learning algorithms and

real-time data analysis to enhance the security and re-

liability of the FASTag ecosystem.

3. (Gunjal et al., 2023) A Survey On FAScam:

FAStag Fraud Detection System: presents a compre-

hensive overview of FAScam, a proposed system de-

signed to detect fraud in the FASTag ecosystem. The

authors discuss the growing prevalence of FASTag

as a digital payment solution for toll collections in

India, along with the associated vulnerabilities that

lead to fraudulent activities such as cloning, unau-

thorized transactions, and manipulation of user ac-

counts. The survey details various machine learning

and data mining techniques that can be employed in

FAScam for effective fraud detection. It highlights

the importance of feature selection and the use of his-

torical transaction data to train models that can dis-

tinguish between legitimate and fraudulent activities.

Fastag Fraud Detection System

905

The paper emphasizes the need for real-time detec-

tion mechanisms that can alert users and operators

of suspicious transactions promptly. Additionally, the

authors evaluate existing fraud detection methodolo-

gies and frameworks in electronic toll collection sys-

tems, pointing out their limitations and proposing en-

hancements through the integration of advanced al-

gorithms. The survey concludes by underscoring the

signiﬁcance of developing a robust, adaptable, and

user-friendly fraud detection system that can not only

protect users but also bolster the overall integrity of

the FASTag system in India’s transportation infras-

tructure.

4. (Roy and Savant, 2022) A Research Paper on

Fast Toll System: provides an in-depth analysis of the

FASTag system, focusing on its operational mecha-

nisms, beneﬁts, and challenges. The authors discuss

how FASTag, an electronic toll collection (ETC) solu-

tion introduced by the National Highways Authority

of India (NHAI), leverages RFID technology to facil-

itate seamless and cashless toll payments. By elim-

inating the need for cash transactions, FASTag sig-

niﬁcantly reduces congestion at toll plazas, enhances

trafﬁc ﬂow, and promotes digital payment adoption.

The paper highlights the various advantages of im-

plementing a FASTag system, including increased ef-

ﬁciency in toll collection, reduced operational costs,

and improved user convenience. It also discusses

the environmental beneﬁts of reduced idle time at

toll booths, which contributes to lower vehicle emis-

sions. However, the authors acknowledge the chal-

lenges faced by the FASTag system, particularly con-

cerning fraud and security vulnerabilities. They em-

phasize the need for effective fraud detection mecha-

nisms to safeguard user accounts and transaction in-

tegrity. The paper calls for further research and devel-

opment in the areas of machine learning and data an-

alytics to enhance the security features of the FASTag

system, ultimately ensuring a more reliable and efﬁ-

cient toll collection process. In conclusion, the au-

thors advocate for continued innovation and improve-

ments in the FASTag infrastructure to address existing

challenges and ensure its successful implementation

in India’s transportation network.

5. (Kumar et al., 2022) FASTag RFID Scam:

examines the vulnerabilities and challenges associ-

ated with the FASTag electronic toll collection sys-

tem in India. The authors provide an overview of

the FASTag system, which utilizes Radio Frequency

Identiﬁcation (RFID) technology to facilitate seam-

less and automated toll payments. While the system

aims to streamline toll collection and reduce conges-

tion at toll plazas, the paper highlights various fraud-

ulent activities that have emerged as a result of its

implementation. The authors discuss different types

of scams related to FASTag, including the cloning

of FASTag devices, unauthorized access to user ac-

counts, and exploitation of system weaknesses. They

emphasize that these fraudulent practices not only

pose ﬁnancial risks to users but also undermine the

overall integrity and trustworthiness of the electronic

toll collection system. To address these issues, the pa-

per calls for the development of robust security mea-

sures and advanced fraud detection mechanisms. The

authors advocate for the application of machine learn-

ing and data analytics to monitor transaction patterns,

detect anomalies, and identify potentially fraudulent

activities in real time. By implementing such so-

lutions, the authors believe that the security of the

FASTag system can be signiﬁcantly enhanced, pro-

viding better protection for users and toll operators

alike. In summary, the paper emphasizes the need for

ongoing vigilance and improvement in the FASTag

system to combat fraud and ensure its success.

6. (Sumangla et al., ) Enhancing Road Safety and

Toll Efﬁciency: provides a comprehensive overview

of the FASTag system and its implications for elec-

tronic toll collection in India. The study begins by de-

tailing the inception and evolution of FASTag, high-

lighting its role in promoting cashless transactions

and streamlining toll payments across the country’s

extensive highway network. By leveraging Radio Fre-

quency Identiﬁcation (RFID) technology, FASTag has

signiﬁcantly reduced waiting times at toll plazas and

improved the overall efﬁciency of the toll collection

process. discusses the beneﬁts of the FASTag system,

including increased transparency, reduced operational

costs, and enhanced convenience for users. The paper

outlines the various features of FASTag, such as real-

time transaction tracking and automatic toll deduc-

tions, which contribute to a more user-friendly expe-

rience. Additionally, the author addresses the govern-

ment’s initiatives to promote the adoption of FASTag

among vehicle owners, emphasizing its mandatory

use for vehicles on national highways. However,

the paper also acknowledges the challenges and lim-

itations associated with the FASTag system. These

include issues related to user awareness, technical

glitches, and the potential for fraud, which can under-

mine user trust in the system. The study emphasizes

the importance of implementing robust security mea-

sures and fraud detection mechanisms to protect users

from potential scams. In conclusion, this paper serves

as a valuable resource for understanding the impact

of FASTag on India’s toll collection landscape. It un-

derscores the need for continuous improvements and

innovations in technology to ensure the sustainability

and effectiveness of electronic toll systems in the fu-

INCOFT 2025 - International Conference on Futuristic Technology

906

ture.

7. (Oza, ) A Descriptive Study on FASTag: Elec-

tronic Toll, Standing Tall: presents an innovative ap-

proach to improving both road safety and toll collec-

tion efﬁciency through the integration of seat belt de-

tection technology with the FASTag billing system.

The authors begin by highlighting the importance of

seat belt use in reducing road accidents and enhancing

passenger safety. They note that despite existing laws

mandating seat belt use, compliance remains a chal-

lenge, leading to higher rates of injuries and fatalities

in vehicle accidents. To address this issue, the paper

proposes a system that automatically detects whether

passengers are wearing seat belts before a vehicle is

allowed to pass through a toll plaza. By combining

this seat belt detection mechanism with the FASTag

system, the authors propose that vehicles not comply-

ing with seat belt regulations could face additional toll

charges or alerts, thereby incentivizing safer driving

behavior. This dual approach aims to reinforce trafﬁc

safety regulations while simultaneously streamlining

the toll collection process. The paper discusses the

technical implementation of this integrated system,

including the use of sensors for seat belt detection

and the necessary modiﬁcations to existing FASTag

infrastructure. The authors provide insights into the

potential beneﬁts of this integration, such as improved

compliance with safety regulations, reduced accident

rates, and enhanced overall efﬁciency at toll plazas.

In conclusion, the paper emphasizes the signiﬁcance

of leveraging technology to create a safer driving en-

vironment while optimizing toll operations. The inte-

gration of seat belt detection with the FASTag system

represents a proactive step towards achieving these

goals, ultimately contributing to better road safety and

efﬁcient toll collection practices.

3 METHODOLOGY

1. Linear Regression: This is used in the pre-

processing stage to set up some sort of baseline pat-

terns and trends in the transactional data. The math-

ematical form of the linear regression can be repre-

sented as:

Y = β

+ β

+ ··· + β

+ ε

Where:

• Y is the dependent variable, e.g., the amount of

the transaction.

• X

, X

, . . . , X

are the independent variables, e.g.,

time and vehicle type.

• β

is the y-intercept.

• β

, β

, . . . , β

are the coefﬁcients of the variables.

• ε is the error term.

Figure 1: Linear Regression

2. Logistic Regression: This is considered the pri-

mary model for fraud detection, using logistic re-

gression as a strategy for binary classiﬁcation, where

each transaction is categorized as either ”fraud” (1) or

”non-fraud” (0). The mathematical model of logistic

regression is presented in the following equation:

P(Y = 1 | X ) =

1 + e

−(β

+β

+···+β

)

Here,

• P(Y = 1 | X) is the probability that a transaction

is fraudulent given the features X.

• e is the base of the natural logarithm.

• β

, β

, . . . , β

are the coefﬁcients estimated during

the training process to improve predictive accu-

racy.

Fastag Fraud Detection System

907

Figure 2: Logistic Regression

Figure 3: Surpervised Machine Learning

4 RESULTS

The results of the given study are actually a fraud de-

tection system of toll transactions through FASTag

data. The system detected some fraudulent trans-

actions based on speciﬁc anomalies or inconsisten-

cies. For example, Transaction 1 may be fraudulent

because of mismatched vehicle dimensions and type

or an amount not paid in the transaction (for exam-

ple, wide discrepancies between what was charged

and what was paid). Transaction 2 may be fraudu-

lent due to a problem like duplicate usage of FASTag

or anomalies in the usage of lane type. Transaction 3

is labeled as ”Not Fraud,” which means there is no ap-

parent anomaly in its associated features. High-speed

variations, missing FASTag IDs, or suspiciously low

transaction amounts may also trigger ﬂags as ob-

served in Transactions 4, 5, 6, 7, and 8. This system

would most probably be using rules combined with

machine learning algorithms to analyze patterns and

deviate from the same to alert about possible fraud.

Figure 4: results

This fraud detection system analyzes multiple fea-

tures like vehicle dimensions, speed, lane type, and

payment discrepancies to detect anomalies in FASTag

transactions.Here we have a transaction amount sig-

niﬁcantly differing from the amount paid. So we are

able to identify fraudulent activity.

Figure 5: results

5 CONCLUSIONS

In conclusion, this paper presents concept of the

FASTag fraud detection system is an effective tool for

identifying anomalies in toll transactions by analysing

various data points, including vehicle details, pay-

ment discrepancies, and behavioral patterns. By

leveraging this system, authorities can signiﬁcantly

reduce revenue leakage, ensure smoother toll opera-

tions, and maintain the integrity of automated toll col-

lection processes. Continuous reﬁnement of the sys-

tem with more advanced machine learning algorithms

and real-time monitoring can further enhance its ac-

curacy and reliability. This approach not only deters

fraudulent activities but also fosters trust among road

users in the digital toll collection system.

Table 1: Performance Matrix of Algorithms

Algorithm Name Performance Matrix

Logistic Regression 0.75

Random Forest 0.75

INCOFT 2025 - International Conference on Futuristic Technology

908

REFERENCES

Bhavar, V., Borkar, V., Hirave, T., Vishwanathan, H.,

and Asati, R. (2023). Fastag fraud detection:

A literature survey. 5th Biennial International

Conference on Nascent Technologies in Engi-

neering (ICNTE).

Eason, G., Noble, B., and Sneddon, I. N. (1955). On

certain integrals of lipschitz-hankel type involv-

ing products of bessel functions. Phil. Trans.

Roy. Soc. London, A247:529–551.

Elissa. Title of paper if known. Unpublished.

Eves, K. and Valasek, J. (2023). Adaptive con-

trol for singularly perturbed systems exam-

ples. Code Ocean. [Online]. Available:

https://codeocean.com/capsule/4989235/tree.

Gunjal, S., Tomy, T., Serpes, A., Yadav, A., and Bo-

dade, V. (2023). A survey on fascam: Fastag

fraud detection system. 5th Biennial Interna-

tional Conference on Nascent Technologies in

Engineering (ICNTE).

Kingma, D. P. and Welling, M. (2013).

Auto-encoding variational bayes.

arXiv:1312.6114. [Online]. Available:

https://arxiv.org/abs/1312.6114.

Kumar, S. R., Prasad, T. S., and N, Z. (2022). Fastag

rﬁd scam. International Journal of Research

Publication and Reviews, 3:23–30.

Liu, S. (2023). Wi-ﬁ energy detection testbed

(12mtc). GitHub repository. [Online].

Available: https://github.com/liustone99/Wi-Fi-

Energy-Detection-Testbed-12MTC.

Nicole, R. Title of paper with only ﬁrst word capital-

ized. J. Name Stand. Abbrev. In press.

Oza, S. M. A descriptive study on fastag: Electronic

toll, standing tall. Unpublished.

Roy, S. and Savant, P. (2022). A research paper on

fast toll system. International Journal for Re-

search in Applied Science & Engineering Tech-

nology (IJRASET), 10:1449–1456.

Sumangla, Prajapat, M., and Kumar, K. Enhancing

road safety and toll efﬁciency: Integrating seat

belt detection with fastag billing at toll plazas.

Unpublished.

U.S. Department of Health and Human Services, Sub-

stance Abuse and Mental Health Services Ad-

ministration, Ofﬁce of Applied Studies (2013).

Treatment episode data set: discharges (teds-d):

concatenated, 2006 to 2009.

Yorozu, Y., Hirano, M., Oka, K., and Tagawa,

Y. (1987). Electron spectroscopy studies on

magneto-optical media and plastic substrate in-

terface. IEEE Transl. J. Magn. Japan, 2:740–

741. [Digests 9th Annual Conf. Magnetics

Japan, p. 301, 1982].

Young, M. (1989). The Technical Writer’s Handbook.

University Science, Mill Valley, CA.

Fastag Fraud Detection System

909