E3S Web of Conferences 399, 05002 (2023) https://doi.org/10.

1051/e3sconf/202339905002
ICONNECT-2023

Intelligent Vehicle Black Box System Using IoT

Sasi Rekha P1, Aadhav A2, Dasharath D2, Karthikeyan S2, Bharath Kumar S2
1
Electrical and Electronics Engineering, M.Kumarasamy College of Engineering, Karur, India
2 Electronics and Instrumentation Engineering, M.Kumarasamy College of Engineering, Karur, India

Abstract. Humans are currently involved in numerous transportation-related accidents on the roads.
Numerous accidents and fatalities caused by human error are the result of the rising number of vehicles on
Indian roads and a lack of enforcement of traffic laws. They also suffer loss of life and valuable property as a
result of those accidents. As a result, accidents involving driver inattention to traffic laws and driver fatigue
follow. The field of computer vision is actively researching real-time eye detection and tracking. Face
alignment can benefit from eye localization and tracking. The duration of eye closure and percentage of eye
closure (PEC) can be used to measure driver fatigue. It is based on a hardware system that uses a camera to
capture images of the driver in real time and software to monitor the driver's eye in order to prevent accidents.
In the event of an accident, GSM and GPS are utilised to track the location of the automobile, and the local
hospital and police are alerted. Thanks to IoT technology, this position may always be found in the cloud
platform service. The 24/7 Governance is notified to call for emergency assistance by using the push and
panic button.

1 Introduction
The Internet of Things (IoT) is a new kind of communication. IoT devices use a variety of wireless
communication protocols to connect to the internet. Machine to machine service is provided to these devices. Devices like
home automation, smart grids, and others are designed to function in real time thanks to the advancing IoT technology. The
potential business is required for the technology's advancement and growth in order to develop the massive platform.
Accidents account for millions of deaths each day, according to the WHO [1]. To find a solution to the problem in
numerous nations, where a vehicle Black Box is being used to find one. Despite numerous campaigns, the problem
continues to worsen on a daily basis, including drunk driving, speeding, and sleep deprivation. The visual data will be
provided by this article, and the cloud will keep track of it with the aid of a low-power microcontroller. This is necessary
because of the constant work that has been done in recent years. The first and second rules of this article are to current data
to end users in a straightforward manner and to envision data by sensing sensors. The automotive technology is primarily
levelled by the automation industry through low power and high range [2].

Fig. 1. Accident data analysis occurs in 2020-21

The black box in the car constantly refreshes the data that collects collision and location tracking information
when it is in active mode. Therefore, the automobile collisions can be easily identified, making it possible for governments


Corresponding author: [email protected]

© The Authors, published by EDP Sciences. This is an open access article distributed under the terms of the Creative Commons
Attribution License 4.0 (https://creativecommons.org/licenses/by/4.0/).
E3S Web of Conferences 399, 05002 (2023) https://doi.org/10.1051/e3sconf/202339905002
ICONNECT-2023

or hospitals to assist the victims. Indian Express reports that there were at least 4,80,652 accidents and 1,50,785 deaths in
India in 2016. The figure suggests that 1,317 road accidents resulted in at least 413 deaths per day. In addition, the data
reveal that 55 accidents per hour during the specified time period resulted in at least 17 fatalities in road accidents. The
Fig.1 below provides the accident data analysis for 2017.

2 Literature survey
Millions of people die annually in traffic accidents, according to statistics from the World Health Organization
(WHO). According to statistics, driver carelessness and fatigue account for the majority of fatal accidents. Nowadays,
urban traffic congestion is getting worse, so some systems like GSM can be used to create a communication unit between
vehicles. Using latitude and longitude to indicate vehicles' proximity, this system aims to monitor vehicles in the
surrounding areas. With the assistance of a message indication, the drivers are notified when these vehicles are extremely
close within this restricted proximity range. Inaccuracies in speed, distance, and response time calculations, among other
issues, are the system's primary drawbacks [3]. According to Y. Ji and S. Wang, it is difficult to identify fatigued driving in
complicated surroundings. In this study, for further exploration of driver tiredness states, a tired driving detection system
based on multi-index fusion and a state recognition network is proposed. The face correction based on the key points of the
eyes, determining the state of the eyes by intercepting a binocular image, and determining the state of the mouth by
intercepting a mouth image based on the left and right corner points are some of the face detection and facial key point
detection techniques used in this study. This makes it possible to detect the driver's head tilt, deflection, and other
movements more precisely [4]. A state recognition network for the eye and a state recognition network for the mouth are
employed in binocular vision to recognise the condition of the mouth. The two characteristics of the mouth state and the
eye state of driver exhaustion are then combined in order to better explore the problem. The method has a detection
accuracy of 98.42 percent on a dataset of public eyes, and a detection accuracy of 97.93 percent on a dataset of open
mouths. In compared to other algorithms already in use, the suggested approach offers the benefits of high accuracy and
straightforward implementation [5].
Abnormal driving behavior, according to J. Hu, L. Xu, and X., may be a major risk to the public and the motorist.
We suggest examining typical driving behaviour in this study to spot aberrant driving. A specialised driver model is
developed to act as the virtual driver and regulate speed using the locally developed neural network and actual Vehicle
Test Data [6]. The driving behaviour is normalised by using the virtual driver to perform the speed-following job required
by a typical driving cycle test, such as the FTP. The typical abnormal driving behaviors of being fatigued or drunk, being
reckless, and using a phone while driving are all characterized and simulated. Based on the abnormality index is developed
and used to quantify the abnormality [7]. To confirm the proposed scheme's efficacy, numerical experiments are carried
out for increasing the rate at which the injured or partially obscured sign is recognized [8]. The PCA method has a very
high rate of accuracy. a recognition self-organizing map (SOM) with a 99 percent hit rate that applied SOM to each level
of RSs. For the purpose of recognizing red traffic signs, our method makes use of SVM with a bagged kernel and RGB
subdivision and form matching detection to cut down on processing time.
The detection algorithm uses grayscale images to make it more resistant to changes in illumination. This situation
prompts the investigation of additional security measures to prevent vehicle thefts. To prevent vehicle theft, we have
decided to implement our project in vehicle safety and security areas. The user can keep track of their vehicle, its routes,
and when it arrives using the Vehicle Tracking System [9]. There is yet another issue. Numerous individuals have chosen
to travel in their own vehicle as a result of recent automobile industry advancements. Vehicle ownership has increased as a
result of this. However, finding a place to park all of these automobiles in major metropolitan areas is a tedious and
challenging task. Towing and other issues arise as a result of improper parking, for which there is currently no solution.
As a result, the parking issue and subsequent towing issue are also addressed in the project. To address all of the
aforementioned issues, the project includes features like keyless door unlocking, keyless ignition control via keypad, and
seatbelt use. The project can also locate a stolen or towed vehicle and address intruder issues from the vehicle's window.
The vehicle's security should be improved by these features [10]. The Black Box Optimization, a more effective and
accurate grouping variant of the differential grouping algorithm, can be proposed as a solution to this issue. By estimating
the magnitude of round-off errors, the DG2 algorithm is used to determine a trustworthy threshold value. However, it is
only used to determine how the two vehicles interact with one another. The test with high-voltage circuit breakers' short-
line fault interruption is utilized by a brand-new black-box arc model [11-12]. During the interruption process, the voltage
and current waveforms are often modelled using black-box arc models. The parameters of this arc model are generated in
this study using an optimization approach over a specified time period in order to minimize the difference between the
simulated and measured arc voltages. These arc characteristics are utilised in a condensed circuit to mimic short-line
failures. Quantitative ratings are given to the model's accuracy in waveform fitting and interruption prediction. As a result,
accidents that take place in a quiet area are not always possible. The following cases have been added to our proposed
system in order to avoid these issues [13-14].

2
E3S Web of Conferences 399, 05002 (2023) https://doi.org/10.1051/e3sconf/202339905002
ICONNECT-2023

3 Proposed system
The suggested solution includes a functioning model of a black box system that may be integrated into a car. The
solution also entails boosting security by guarding against harm to the black box data. The BLXDAQ (BlackboX Data
Acquisition) module stores data on a vehicle's movement parameter, primarily for the purposes of accident analysis and
safety measures. The sensors that are incorporated into the engine and system of the vehicle are connected to these
modules. The vehicle's location in the cloud is updated using the IOT module. It indicates the vehicles' precise latitude and
longitude in Fig.2.

Fig. 2. Proposed diagram of Black-Box system

Due to the increasing movement of heavy vehicles, driving at night has become a challenging situation marked by
numerous accidents and concerns for both transportation authorities and the general public. After a few days of continuous
driving, the drivers' reduced reflexes reduce their ability to drive, resulting in accidents. The drivers are required to drive
with little rest. Tiredness is found to be the cause of most accidents where people fall asleep. A combination of symptoms
like poor performance and a subjective sense of drowsiness is known as fatigue. There is still no widely accepted definition
for the term "fatigue" despite the extensive research that has been carried out. There are various types of fatigue, including
the ones outlined below, which are viewed from the perspective of the individual organs' functionality: 1) physical fatigue
in a specific area (such as a skeletal or ocular muscle); 2) general physical exhaustion as a result of extensive manual labor;
3) sleepiness (fatigue of the central nervous system); 4) mental exhaustion (inability to do anything). We can use linear
discriminate analysis to notice faces and track the states of the eyes with greater precision in the proposed system. The
carter's eyes will be closed as a corrective measure in the event of abnormal behavior, and an alarm will sound. After
correctly locating the driver's head and eyes in the camera image, the system moves on to the analysis phase. For the
purpose of detecting drowsiness, various Image Processing techniques are used to preprocess this image. Finally, provide a
voice, SMS, and email alert admin with face recognition alert system. Additionally, an embedded system is used to slow
the vehicle down to keep the driver from getting into an accident.

3
E3S Web of Conferences 399, 05002 (2023) https://doi.org/10.1051/e3sconf/202339905002
ICONNECT-2023

3.1 Microcontroller
The development processes determine a significant portion of the final system's functionality, therefore effort is
made to ensure that the best techniques, resources, and tools are employed. IDE will be used to develop the system. There
will also be usage of embedded C programming. With the purpose of enhancing performance, several system components
will be migrated to Android.

3.2 PSU Board

This board is linked to the relay chip, which is then linked to the system for the vehicle model. The
microcontroller, DC motor, and sensors' power needs are mostly met by this board.

3.3 Buzzer

This driver, which is powered by the microcontroller board, supplies electricity to the buzzer. To supply power to
the buzzer, the enable pins of this driver must be high.

3.4 Camera
To capture continuous photographs of the traffic signs and signals in the actual world, a camera is employed. We
can transmit the PC the photos obtained from the camera so it can operate the automobile based on those images.

3.5 Python software

Python is a high-level, all-purpose programming language that is often used for making projects of various kinds, including
those for business and academic endeavours. Many versions of its programme are available, including IDLE, Python 2, and
Python 3. Moreover, there are additional versions of Python IDLE for python programming accessible within these two
categories.

3.6 Open CV
Open-Source Computer Vision is what it stands for. Its programming function library is mostly for real-time
computer visions. It is mostly used for image processing, where it is employed for tasks such as face recognition, object
detection, picture recognition, traces, and other related ones. It offers more than 2500 optimise algorithms for a variety of
conventional algorithms and cutting-edge computer vision methods.

4 Intelligent vehicle black box system

4.1 Working principle

The majority of real-time apps failed due to insufficient power usage. The controller of this Intelligent Vehicle
Black Box employing IoT utilizes less electricity and prefers real-time applications. Several sensors, such as
accelerometers, ultrasonic sensors, and breathalyzers, are used to ensure the driver's safety. When a motorist exceeds his or
her alcohol consumption limitations, the owner or the vehicle authority's designated contact is instantly alerted. By
utilizing the technology, the driver will be vigilant and will refrain from such activities. When a collision is near, the
ultrasonic sensor warns the driver by keeping track of the distance between the surrounding automobiles.
After a collision, the accelerometer measures the tilt values and simultaneously transmits the message and
position through GSM and GPS. The panic button is used to contact emergency contacts, whereas the push button is
intended to request aid from the government. Sensor readings and position are sent to the cloud for tracking reasons.

4.2 Hardware Design

The hardware is now integrated into automobiles as IoT technology develops. These vehicles' sensors gather
information and send it to the microcontroller. As soon as the accident happens or the driver drinks, the emergency
contacts will be notified. Vehicles consume a lot of energy, hence ESP32 is utilised to keep the programme running for a
longer duration.

4
E3S Web of Conferences 399, 05002 (2023) https://doi.org/10.1051/e3sconf/202339905002
ICONNECT-2023

4.3 Designing Operational Flow

The operational flow design is shown in Fig.3. The many states in which the system transitions while the intended
action is being performed are described by the following steps:
• Step-1: The alcohol sensor, ultrasonic sensor, and accelerometer start initializing after receiving power from the
system.
• Step-2: GPS is always used to transmit the location to the cloud.
• Step 3: The ultrasonic sensors located in the front and rear of the car alert the driver through a buzzer if they
assess that the space between the other vehicles in the area is narrowing.
• Step 4: Now, if these sensors exceed the limit, proceed to Step 4. If not, proceed to Step 3.
• Step 5: The message and the location are sent right away if the accelerometer detects a collision.
• Step 6: After the controller receives this message, it performs three tasks simultaneously. The message and the
location will be delivered to the supplied cell phone number using a GSM module. The Cloud stores this location.

The system is put into a moving car once all the parts are connected correctly. The vibration sensor detects the acceleration
caused by the collision in the event of an accident, and its ADC output pins are directly connected to Arduino ADC pin A0.
The system is successfully initialized, the sensor is calibrated, and the samples of the sensor data vehicle are displayed on
the LCD after the program has been successfully compiled and uploaded on the Arduino IDE. After receiving a GPS
signal, the vehicle's current position's latitude and longitude, as well as its speed, are shown. When the accelerometer is
shook unnaturally, such as in the case of an accident with a quick shift in axis, the latitude and longitude are also
communicated through Google maps. Moreover, an SMS is delivered to the phone number in the code. The message is
received at the specified location and mobile phone number.

Fig. 3 Proposed Circuit Diagram

5 Conclusion and future scope

To solve global concerns, the Internet of Things (IoT) technology of today is quickly developing. The alerting of
the driver in collision situations is the primary focus of this paper, and the location can be easily tracked by utilizing Cloud
Computing Services. We offered a low-power microcontroller that could act as the device's main controller during
automation in the hardware implementation as our contribution. We are certain that with major assistance from embedded
systems, the Internet of Things, and cloud computing, the Intelligent Vehicle Black Box leveraging IoT will be dependable
and power efficient in real-time applications. In the future, power consumption can be greatly decreased by GPS and GSM
modules, which can be readily added into hardware, have a rechargeable battery, and can operate the device for a longer
amount of time. The suggested system's objective is to offer details regarding the accident's location. Receiving help and

5
E3S Web of Conferences 399, 05002 (2023) https://doi.org/10.1051/e3sconf/202339905002
ICONNECT-2023

aid for the accident victim is made simpler. This system's GPS module is utilised to find the car. To deliver accident
information, GSM is employed. The suggested systems deliver results that are adequate.

References

1. G. Hayes, F. Blosser, “Motor Vehicle Crashes Claim More than a Million Lives Worldwide”, CDC Injury Center
Media Relations, Press Release, April, (2004).

2. Abdallah Kassem, Rabih Jabr, Ghady Salamouni, Ziad Khairallah Maalouf, “Vehicle Black Box System”,
Proceedings of the 2nd Annual IEEE System Conference, IEEE (2008).

3. Y. Ji, S. Wang, Y. Zhao, J. Wei, and Y. Lu, ‘‘Fatigue state detection based on multi-index fusion and state
recognition network,’’ IEEE Access, 7 (2019).

4. C Bhuvaneshwari, A Manjunathan, “Reimbursement of sensor nodes and path optimization”, Materials Today:
Proceedings, 45 (2021).
5. Koichi Fujiwara, Erika Abe, Keisuke Kamat, Chikao Nakayama, Yoko Suzuki, Toshitaka Yamakawa, “Heart Rate
Variability-based Driver Drowsiness Detection and its Validation with EEG,” IEEE Transactions on Bio. Engg. 66, 6,
(2019).

6. Bhuvaneshwari C, Manjunathan A, “Advanced gesture recognition system using long-term recurrent convolution
network”, Materials Today Proceedings, 21 (2020).
7. S. Mühlbacher-Karrer et al., “A Driver State Detection System— Combining a Capacitive Hand Detection Sensor
with Physiological Sensors,” IEEE Trans. Instrum. Meas., 66, 4 (2017).

8. K. Fujiwara et al., “Epileptic seizure prediction based on multivariate statistical process control of heart rate
variability features,” IEEE Trans. Biomed. Eng., 63, 6 (2016).

9. A Manjunathan, AS Kumar, S Udhayanan, CT Selvi, AA Stonier, “Design of Autonomous Vehicle Control using
IoT”, IOP Conference Series: Materials Science and Engineering, 1055, 1, 012008, (2021).
10. B. K. Dar, M. A. Shah, S. U. Islam, C. Maple, S. Mussadiq, and S. Khan, “Delay-Aware Accident Detection and
Response System Using Fog Computing,” IEEE Access, 7 (2019)

11. Manjunathan Alagarsamy, Prabakaran Kasinathan, Geethalakshmi Manickam, Prabu Ragavendiran Duruvarajan,
Jeevitha Sakkarai, Kannadhasan Suriyan, “IoT based E-vehicle monitoring system using sensors and imaging
processing algorithm”, International Journal of Reconfigurable and Embedded Systems, 11, 2 (2022).
12. C Ambika Bhuvaneswari, ED Kanmani Ruby, A Manjunathan, R Balamurugan, P Jenopaul, Belachew Zegale Tizazu,
Effects of Novel Material Field Effect Transistor for Heterogeneous Energy and Traffic-Aware Secure Applications”,
Advances in Materials Science and Engineering, Article ID 9085854 (2021).

13. Rajagopal Sudarmani, Kanagaraj Venusamy, Sathish Sivaraman, Poongodi Jayaraman, Kannadhasan Suriyan,
Manjunathan Alagarsamy, “Machine to machine communication enabled internet of things: a review”, International
Journal of Reconfigurable and Embedded Systems, 11, 2, (2022).

14. G Gabriel Santhosh Kumar, A Manjunathan, T Sathesh Kumar, M Mathankumar, G Anushree, N Mohanasundaram,
“Analysis on the Impacts of Light Emitting Diodes in Power System using FLUKE 430-II D Power Quality Analyser”,
International Conference on Advancements in Electrical, Electronics, Communication, Computing and Automation
(ICAECA), 08-09, October 2021, Coimbatore, India (2021).