Automatic Driver Drowsiness Detection System

Ponnuviji N P Varsha E Vindya Vasini D

Department of Computer Department of Computer Department of Computer
Science and Engineering Science And Engineering Science and Engineering
RMK College of Engineering RMK College of Engineering RMK College of Engineering
And Technology And Technology and Technology
Thiruvallur, India Thiruvallur, India Thiruvallur, India

driver drowsiness in order to reduce collisions and enhance

highway safety. A camera pointed directly at the driver's face
and captured for a real- time image is one way that has been
created to detect drowsiness in drivers. The apparatus has the
Abstract—The prediction of a driver's sleepiness capability to observe eyes and ascertain if they are closed or
prevention system uses the latest advancements in open. When drowsiness is detected in such a situation, the
artificial intelligence and technology to effectively handle driver is alerted with a warning signal. The percentage of a
driver weariness as a separate cause of traffic accidents. time interval that the eye is closed can be used to calculate
One of the primary reasons of accidents most of the time the time interval of eye closure. The system determines that
is fatigue. Drivers typically become drowsy when they are the driver is about to fall asleep if their cumulative eye
stressed, sleep deprived, intoxicated, have an incorrect closure time exceeds a predetermined threshold. In such case,
sitting posture when driving, or experience numerous an alarm will ring to notify the driver.
other sleep-related issues, such as disorders. Over the past
few decades, advances in artificial intelligence and II. LITERATURE SURVEY
technology have resulted in enhancements to driver [1] Automated Identification of Sleepiness in EEG Data via
monitoring systems. Create a system to identify driver Time Analysis Over 20% of incidents resulting in injuries and
weariness and stop accidents caused by it. Apple has fatalities worldwide have been linked to driver fatigue. An
identified a Python application for detecting driver inebriated driver performs worse and is less adept at handling
fatigue. Therefore, this proposed model uses the driver’s a car. The goal of this study is to automatically identify the
face position to detect whether he/she is falling asleep or degree of tiredness in EEG recordings using temporal
not and an alarm will be sounded if he/she needs it. When analysis techniques. The EEG data were divided into 5-
it senses that you are drowsing with even though you are second intervals after being subjected to band-pass filtering
driving it lowers your vehicle speed gradually with with a cut-off frequency of 0.5 Hz to 60 Hz to eliminate
parking light simultaneously turned on. noise.One EEG channel was used to compute three features
Keywords—convolution neural network, feature extraction,
that allowed the stages of alertness and fatigue to be
image pre-processing, ESP32
distinguished.Up to 85.7% of correct detections of alertness
I. INTRODUCTION and drowsiness are achieved using this method. Based on the
parameters and the data, both stages can be identified. The
We suggest a Drowsiness Detection System in this project to ability to acquire real-time outputs makes it simply accessible.
enhance personal security, vehicle security and safety. The These parameters can be used by an autonomous sleepiness
purpose of this research is to develop a prototype drowsiness detection system in cars, reducing the danger associated with
detecting system. The main goal is to create a system that can drowsy driving. The MITBIH Polysomnographic Database
precisely determine if a driver's eyes are open or closed in on physionet.org provided examples for testing the algorithm
real time. It is thought that symptoms of driver weariness can proposed here. AUTHORS: Amit Prasad, Shivansh Jagga,
be detected early enough to prevent an accident by keeping a and Ishita Dey.[2] An Identifying Driver Drowsiness System
watch on the eyes. To detect tiredness, a pattern of facial Based on Vision Driver fatigue is a primary contributing
images is employed in conjunction with tracking eye factor in many traffic incidents.The number of fatalities and
movements and blink frequency. Face image analysis is a injuries rises annually on a global scale. This paper presents
significant area of research with applications in human an Advanced Driver Assistance System (ADAS) module that
identification, virtual tools, and facial recognition. uses artificial intelligence and visual information to detect
Applications for face analysis include virtual tools, security driver drowsiness automatically, reducing the number of
systems for human identification, and facial recognition, accidents caused by fatigued drivers and improving
making it a popular field of study. This project uses the transportation safety. We proposed a method to locate, track,
localization of the eyes, which entails analyzing a face image and analyze the driver's face and eyes to quantify PERCLOS,
and using a Python software to determine the position of the a verified scientific sleepy indication associated with slow
eyes. After locating the eyes's position, the technique eye closure.Belal Alshaqaqi, Mohamed El Amine Ouis,
Meriem Boumehed, and Abdullah Salem Baquhaizel are the
authors.[3]An analysis of a smart car's automatic multimodal
sleepiness detection system-- As artificial intelligence and
technology become more prevalent in the automotive sector,
is intended to identify tiredness and ascertain whether the driver drowsiness monitoring systems have attracted a lot of
eyes are open or closed. The goal of this research is to identify interest as a way to improve safety and reduce fatigue-related

XXX-X-XXXX-XXXX-X/XX/$XX.00 ©20XX IEEE

accidents.However, these technologies miss the fact that Emocds dataset, which is accessible to the public. Our
every driver exhibits a unique collection of physiological and approach outperformed previously reported approaches by a
behavioral signs of fatigue, making objective evaluation of little margin. The entire network was a time-distributed layer
drowsiness a challenging task.. The AUTOMOTIVE project of the CNN structure in its entirety. The suggested study
investigated the use of immersive driving simulators to demonstrated improved accuracy and yielded general
enable driver-specific sleepiness detection in smart cars findings by taking into account different conditions. Datasets
through the application of machine learning and signal as YawDD, NTHU-DDD, and the proposed EMOCDS (Eye
processing techniques. In general, large-scale research on and Mouth Open Close Data Set) were used to evaluate it.
biometrics, such as the face and electrocardiogram (ECG), The driver's level of fatigue is determined by Python image
enables continual learning of drowsiness models unique to processing. The micro-controller activates the buzzer alert
each subject for more efficient monitoring.This article's goal whenever it senses drowsiness. Immediately switch off the
is to provide a comprehensive and all-encompassing engine, activate the LED indicator, and update the IOT
overview of the research and development work done for the website with the current position.
AUTOMOTIVE project. It covers the various topics covered
A. Description of the Proposed System
and explains how it ultimately advances our goal of
improving driver drowsiness monitoring.Pedro Ferreira, Joao
Rebeiro Pinto, and Telma Esteves are the authors.[4] A 1) ESP32- is a highly reasonably priced System on Chip
Driver Who Operates Automatically GSM-based drowsiness (SoC) Micro controller manufactured by Espressif Systems,
alarm system-- The majority of road accidents are mostly the same company that produced the popular ESP8266 SoC.
caused by drowsy drivers.The fact that thousands of cars are The ESP8266 SoC is replaced by the 32-bit Tensilica Xtensa
driven on the roads every day makes it difficult to manually LX6 Microprocessor, which has integrated Bluetooth and
track down a sleepy driver. Thus, every car must have a Wi-Fi and may operate on single or dual cores. The ESP32
system that automatically stops the vehicle if it senses that the shares the same benefits as the ESP8266, including inbuilt RF
driver is tired. Furthermore, in the event that the driver falls balun, filters, antenna switch, low-noise receive amplifier,
asleep, the car will stop and the driver's temperature, heart and power amplifier. Because of this, developing hardware
rate, and respiration rate will be recorded and shown on the based on the ESP32 is quite easy because it needs
LCD. These three metrics are crucial since they indicate the comparatively few external components.
driver's physical condition. These settings are closely 2) ADAPTER- An adapter is a device that modifies the
examined, and in the event of an emergency, the ward properties of an electrical system or gadget so that it can
supervisor will get in touch with the physician.Better real- function with one that wouldn't normally work with it. While
time eye state detection technology to recognize drivers' some only modify an electrical connector's external design,
drowsiness One of the main issues with driver fatigue others improve its transmission or power properties.
identification is real-time eye state detection. In order to Typically, an adapter in a computer is integrated into a card
detect driver drowsiness, a real-time eye condition that fits into a motherboard slot. The card modifies data that
monitoring system is proposed in this study. is transferred between the microprocessor of the computer
and the supported peripherals..
To meet the requirements of the drowsiness detection 3) RELAY- An electro-mechanical switch, or relay, is
methods, the system optimizes a number of image processing activated by an electric current. A four-relay board
algorithms. Initially, the enhanced Haar-like feature detection arrangement has three circuits: one for the driver, one for the
method is employed to locate the facial region. The power supply, and one for isolation. Relays are used in the
geometrical location of the eye on the face and the horizontal assembly of that circuit. Switching operations in the driving
projection of the detected face are then applied to obtain the circuit are carried out by transistors.The relay is switched via
eye area. And lastly, a novel complexity function for the transistor. Reverse voltage from the relay is prevented via
determining the eye state that has a dynamic threshold. an isolation circuit, shielding the transistor and controller
Compared to many other approaches, the strategy described from harm. The micro-c provides the input pulse needed to
in this study strikes a superior compromise between accuracy flip the transistor.
and efficiency. Based on the results of the experiment, the 4)DC MOTOR- Geared DC motors can be conceptualized as
system, modified using Intel IPP (Integrated Performance an extension of DC motors. A geared DC motor has a gear
Primitives), might be able to meet the requirements for real- system attached to the motor. RPM, or shaft rotations per
time acquisition. minute, is the unit of measurement used to describe the
II. SYSTEM DESCRIPTION motor's speed. Reducing speed and increasing torque are
In this integrated technique, the Eye and Mouth Closure facilitated by the gear assembly. The correct set of gears can
Status (PERCLOS) is determined along with the newly be used to adjust the speed at which a gear motor operates.
proposed vector FAR (Facial Aspect Ratio), similar to EAR The concept of gear reduction holds that by utilizing gears, a
and MAR. vehicle's torque can be raised while its speed can be
This makes it easier to determine whether someone is decreased. A DC motor can be run at a voltage that is less
yawning or has their eyes closed. It also helps to identify any than its rated voltage. Below 1000 rpm, the motor will not run
frames that show hand actions used by people to try to control smoothly and the speed will become erratic.
their sleepiness, such as nodding or covering an open mouth 5) LCD- This LCD screen was designed with E-blocks in
with a hand. building a deep learning model that performed mind. An LCD display with 16 characters and 2 lines of
well in comparison to earlier research using the Drowsy alphanumeric characters is connected to a single 9-way D-
type connector. Consequently, the majority of E-Block I/O ⚫ After bringing it to stop, it activates parking light so as to
ports can be connected to the device. The information about alert other drivers about pressing on brakes
the serial data format required for the LCD display can be
found in the user manual that follows. The display also
requires a 5V power supply. Please take care not to exceed ⚫ Also, buzzer sound is produced by this system while
5V as this could damage the device. The E-blocks The best displaying an alert message on car mirror via LCD
display.
sources for generating 5V are multiprogrammers or fixed
regulated power supplies with a 5V output. The 16 x 2
intelligent alphanumeric dot matrix displays may show 224
⚫ Implementing this system aims at promptly intervening
different characters and symbols.
when fatigue signs are detected in order to reduce
6) UART- The UART receives data from a data bus and instances of accidents due to slow response. Real-time
transfers it. Data can be transferred from another device, such monitoring, reference image comparison and automatic
as a CPU, RAM, or microcontroller, to the UART via the data responses are combined altogether for improved road
bus. Data is exchanged in parallel between the transmitting safety generally leading to lower risks from drowsy
UART and the data bus. The transmitting UART creates the driving.
data packet by adding a start bit, a parity bit, and a stop bit
after receiving the parallel data from the data bus. After that,
the data packet is sent bit by bit, serially, to the Tx pin. Bit by III. SOFTWARE DESCRIPTION
bit, the data packet is read by the receiving UART at its Rx
pin. The start bit, parity bit, and stop bit are then eliminated
by the receiving UART before the data is used again in 1)ARDUINO SOFTWARE IDE - The Arduino Software
parallel fashion. Lastly, information transmitted by UART is (IDE) includes a text editor for writing code, a toolbar with
arranged into packets and transferred via the receiving UART buttons for frequently used activities, a message box, a text
in parallel to the data bus on the receiving end. One start bit, console, and other menus. To upload and interact with
one or two stop bits, five to nine data bits (depending on the programs, it establishes a connection with the Arduino
UART), and an optional parity bit make up each packet. hardware. Sketches are programs created using the Arduino
Software (IDE). Using a text editor, these images were
created and saved as files with the.ino extension. The editor
offers functions for text replacement and search in addition to
cutting and copying. When exporting and saving, the message
box provides feedback in addition to displaying faults.
Complete error warnings as well as text output from the
Arduino Software (IDE) are displayed in the terminal.

2)EMBEDDED C- Programming in high-level languages

has long been a part of embedded system development.
However, a lot of systems still require assembly
programming, particularly those that employ digital-signal
processors (DSPs). Assembly language programmers that are
well-versed in processor architecture frequently write code
for DSPs. Performance is the main driving force behind this
approach, even with assembly programming's drawbacks
over high-level language programming. If, for instance, video
decoding consumes 80% of the available CPU cycles instead
Fig.1Block schematic of the suggested setup of 90%, then twice as many CPU cycles are available for
audio processing. The performance of DSP processors is
B. Working principle coupled to end-user features in many real-time applications.
⚫ The application depicted herein is a real-time drowsiness For signal processing, SPs need a highly specialized
detection system that has been designed to improve architecture in order to meet performance requirements.
driver’s safety and prevent possible accidents caused by Applications that don't exceed the price and power
sleepy driving. This system consists of two parts: an constraints placed by consumer applications. In contrast to
initial configuration stage where the consumer uploads traditional RISC architectures, DSPs provide a data channel
two pictures – one with open eyes and another with that feeds directly into the arithmetic units from memory-
closed eyes, as well as a real-time monitoring stage using access units. The address registers are moved to separate
a digital camera which is mounted beside the register files close to the memory units after being removed
speedometer in the car.The python code compares the from the general-purpose register file.
position of the driver’s eye to that of the camera. If it
detects that for some time his/her eyes have been closed, 3)PYTHON- SPs require a highly specialized architecture
there is a case for drowsiness and thus gradually slowing
for signal processing in order to meet performance
down and ultimately stopping the vehicle.
requirements. Applications that respect the limitations
imposed by consumer applications on price and power. DSPs
offer a data channel that feeds directly into the arithmetic
units from memory-access units, in contrast to conventional security and support, leading to timely responses in
RISC systems. The address registers are moved to separate cases of accidents involving tired drivers. This extra
register files close to the memory units after being removed feature notifies the driver's next of kin and notifies the
from the general-purpose register file. nearest police station of the driver's location and
condition.
IV.RESULTS
The suggested concept has been put into practice, and a
prototype has been created to evaluate its VI. CONCLUSION
effectiveness.1.LED;2.DC motor;3.buzzer;4.LCD;5.ESP32
The system is designed to track the eye and to identify fatigue
3 as a means of detecting sleepiness, using advanced image
3 processing algorithms. The posture of the eyes in this case
can be accurately identified by the system with high certainty
and dependability through use of self-built image processing
techniques. In addition, it uses image processing to extract
relevant features from the captured images of a driver’s face
focusing on region around a driver’s eyes. It then uses
4 intricate algorithms to analyze these features in order to
4 deduce how alert the driver is. One main advantage is that it
5 is diagnostic in nature since it does not need any interaction
with drivers physically or application of invasive sensors.
Instead, just pictures taken by a camera are sufficient for the
system thus causing minimal discomfort and disturbance to
the driver. Moreover, employing image processing has
several advantages including highly accurate and reliable
detection systems for signs indicating drowsiness. By
22 analyzing minute changes in eye behaviors and movements,
early indications about fatigue may be given by this method
which will help prevent possible accidents before they
happen. Thus overall, its diagnostic nature coupled with its
1 applicability makes this system an invaluable tool for
1 improving Overall, the diagnostic nature and the
effectiveness of image processing make this system a
valuable tool for enhancing safety on the roads by detecting
drowsiness without causing annoyance or interference to the
driver.
V. FUTURE ENHANCEMENT
REFERENCES
A few potential future developments are as follows:
[1] FARS Data-Tables, National Highway Traffic Safety Administration
⚫ One highly valued future direction is adding Internet of (2011).
Technology (IOT) as an extra function to the drowsiness
[2] Office Safety Integration, Federal Highway Administration, 2010.
identification system. IOT enables the absolute
transmission of real-time driver position and status data [3] Onboard Measurement and Warning module for irregular vehicle
to designated recipients, such as the driver's family behavior, Tang-Hsien Chang, Chih-Sheng Hsu, Chieh Wang, Li-Kai Yang
members and the closest police station. This allows Intelligent Transportation Systems Transactions, IEEE
(2008).
remote communications with the system and allows for
continual monitoring of it. Should the level of fatigue [4] "Automated Video-based Measurement of Eye Closure for Detecting
detection surpass a predetermined threshold, Behavioral Microsleep" by A. Malla, P. Davidson, P. Bones, R. Green, and
information about the subject's current state and location R. Jones
can be automatically transmitted through IOT-enabled [5]. P. Viola and M. Jones, "Accelerated Object Identification through an
communication channels to certain recipients. Enhanced Sequence of Basic Characteristics"

⚫ With this information, officials at the closest police [6]"Driver status monitoring systems for smart vehicles using physiological
sensors," by Y. Choi, S. I. Han, S.-H. Kong, and H. Ko
station will be able to offer expert assistance or
interventions as soon as needed, in order to give prompt [8] "Intelligent safety warning and alert system for car driving," S.-N. Lu,
feedback. In a similar vein, informing the deceased's H.-W. Tseng, Y.-H. Lee, Y.-G. Jan, and W.-C. Lee J. 2010.
close relatives of their situation would prompt them to [9] "On-road and simulated driving," by D. R. Mayhew, H. M. Simpson,
take appropriate action or perhaps extend assistance. K. M. Wood, L. Lonero, K. M. Clinton, and A. G. Johnson

⚫ Integration of IOT technology often provides

drowsiness detection systems with an extra degree of
[10] "Visual Analysis of Eye State and Head Pose for Driver Alertness
Monitoring," Ralph Oyini M. Bouna, Seong G. Kong

[11] Arun Sahayadhas and Kenneth Sundarraj; An Assessment of Detecting

Driver Drowsiness Using Sensors; Malaysia, 2012.

[12] Koji Oguri, Haruhiko Hashimoto, Keitarou Ishihara, and Katsuki

Hayashi. Individualized Drowsiness Detection by Pulse Wave Analysis
using Neural Network while Driving. The 8th International Conference
Proceedings.
[13] Xiaomin Lian, Jie Ding, Diange Yang, and Keqiang Li. Current state [
and trends with operator fatigue methods.