ISSN (e): 2250 – 3005 || Volume, 09 || Issue, 6 || June– 2019 ||

International Journal of Computational Engineering Research (IJCER)

Intelligent Road Accidents Avoidance: A Review

1
Noorishta Hashmi, Dr.M Akheela Khanum
1
M.Tech Scholar, Department of Computer Science, Integral University, Lucknow, India.
2
Dr.M Akheela Khanum, Department of Computer Science, Integral University, Lucknow, India.
Corresponding Author: Noorishta Hashmi

ABSTRACT
With the increase in population and so also the number of vehicles on the roads there have been an
increase in the number of road accidents taking place daily. This era is called as the era of Machines
and Artificial Intelligence. It would be worthy if some intelligent techniques are used to avoid road
accidents thereby protecting the human lives from accidental deaths. Literature shows a number of such
techniques are in place which makes use of one or more parameters to predict the likelihood of accident
and alerts the driver. In this paper we have reviewed many such papers to gain insight into the state-of-
the-art and find out areas of improvements in those techniques. The results of the literature review
indicate that many techniques for avoiding road accidents are proposed. Some makes use of data
mining, some makes use of IoT, Cloud, and Neural networks. Various parameters such as driving
pattern, speed and drowsiness are used to predict the likelihood of accidents.
Keywords: Driving behavior, Patterns, Drowsiness.

----------------------------------------------------------------------------------------------------------------------------- ----------
Date of Submission: 26-05-2019 Date of acceptance:08-06-2019
----------------------------------------------------------------------------------------------------------------------------- ----------

I. INTRODUCTION
Deaths due to road accidents have become a common phenomenon nowadays. As the rate of the population is
growing likewise builds the rate of road accidents. Main causes behind these road accidents are lack of training
institutes, unskilled drivers, poor road conditions, use of cell phone during driving, consuming alcohol while
driving; overloading. As per the World Health Organization (WHO), street crashes kill 1.2 million. [1]
Throughout the most recent decade, street crash has turned into the tenth driving reason for death on the planet,
and is predicted to rise to the fifth position by 2030. India is the main supporter of worldwide street crash
mortality . In the most recent decade alone, India lost 1.3 million individuals to street crashes and another 5.3
million were debilitated forever. In the most recent decade alone, India lost 1.3 million individuals to street
crashes and another 5.3 million have been truly harmed [2]. One serious road accident in the country occurs
every minute and 16 die on Indian roads every hour.1214 road crashes occur every day in India. Two wheelers
account for 25% of total road crash deaths.20 children under the age of 14 die every day due to road crashes in
in the country.377 people die every day, equivalent to a jumbo jet crashing every day[3]. Two people die every
hour in Uttar Pradesh – State with maximum number of road crash deaths Various Techniques are being
introduced to reduce accidents. Many monitoring methods are come to analyze Driver behavior. Drowsiness is
recognized as an important factor in a vehicle accident. In recent years, driver drowsiness and distraction have
been important factors in a large number of accidents because they reduce driver perception level and decision-
making capability, which negatively affect the ability to control the vehicle. One way to reduce these kinds of
accidents would be through monitoring driver and driving behavior and alerting the driver when they are drowsy
or in a distracted state.

II. LITERATURE REVIEW

Several research papers in the area of road accident avoidance were reviewed systematically. Some of the most
relevant researches have been briefed in this section.
In this paper [4], a D&R sense System is proposed that helps in categorizing the driving style of drivers, access
the road quality and also warned the driver in real time to make safe driving with the help of GPS, accelerometer
and Smartphone sensor. An algorithm named SVM is used to do pattern recognition into different categories
such as aggressive or calm driving, bumpy roads, etc. The result shows that the locally running fast DTW gives
an accuracy of 86.36% and SVM gives 95.45% but when the same thing was tested on a threshold-based
algorithm, accuracy decreases dramatically.
In this paper [5], the author presents an overview of different studies on monitoring $ detection of drowsiness
and micro sleep (MS) while driving. At first, a framework is proposed for utilizing methods of pattern

www.ijceronline.com Open Access Journal Page 35

 Intelligent Road Accidents Avoidance: A Review

recognition $ of computational intelligence in order to detect $ provide AS. Secondly, comparisons of different
biosignals are done to their values for MS detection. Thirdly, comparisons algorithm are done and also the pros
and cons of SVM for this task. Fourthly, investigation of automatic relevance determination introduced. Fifthly
different problems of inter and intra individual variability are addressed. Sixthly, By learning from data, the MS
detection algorithm has been established which should be kept in mind. Seventhly, a new measure of drowsiness
$strong central fatigue is presented which is known as MS density. Eighthly, the method to evaluate commercial
fatigue monitoring terminology ( EMT) is presented in the form of applications.
In this paper [6], the author proposed an assistance strategy for safe driving which mainly aims for the normality
of driver's behavior. An analysis is done on the driver's behavior using steering wheel operation. Fundamental
experiments are done with the help of driving simulators system validate the proposed system. Results show that
80% of the subject drivers are not aware of their driving style.
In this paper [7], the author checks the possibilities of automatic detection of the distraction of driver due to lots
of cognitive tasks from driving behavior was studied and with the help of seveors data, a recording of a driving
simulator which was further accessed for analysis. For a cognitive test, the driver was asked to engage in
conversation with the co-passengers and also do verbal observed that there was a big difference in terms of a
statistical test in driving behaviour for action driving and distracted driving. The difference for an inexperienced
driver was critically high whereas or expert drivers it is low.
In this paper, [8] presents a real-time lane detection and driver fatigue or driver drowsiness detection system,
which can effectively detect the anomaly while driving. The main concentration of the author on avoiding the
road accident by concentrating on drunk drivers or drowsiness. Firstly lane detection using Hough Transform
and secondly Eye detection of the driver is done. Blinking time of the driver are used. He closes his eyes more
than 10 to 15 sec an alarm showing alert will glow. The image segmentation is done, Otsu thresholding is
performed and Canny edge detection is done, the result obtained is then applied with Circle detection Hough
Transform, to detect the eyes. For lane, detection is done if the driver changes the lane without signal than the
system takes an error and sends the warning sign. The webcam is used to detect lane and to monitor face. This
system will be particularly useful for drivers travelling on long routes, night drivers and for people who drink
and drive.
In this paper [9], the author predicts driving behavior through analyzing driver Gaze pattern. Our experiments
suggest that a driver's gaze patterns appear prior to and correlate with the driving behaviors for driving behavior
prediction. We accordingly propose GazMon, an active driving behavior monitoring and prediction framework
for driving assistance applications. CNN and LSTM algorithm are used with cuDNN on Nvidia
GTX1050TiGPUS. Nvidia surrounds technology enables to combine display to create the most combine display
to create the immersive emulating. In this Tobi eye X 4c3 as use for eye tracking device. CNN gives 22% hi
accuracy LSTM gives 36% and 40 % accuracy with SVM.
In this paper, [10], the author focused on the application layer in the cloud computing platform. Python has
adopted in this to calculate the Kurtosis and skewness in statistics of each driving route. That decision tree,
classification technique is used to generate the analyzing knowledge for driver behaviour analysis. At last
summarized from completed the decision tree whether the drive is defensive, weak defensive, aggressive, and
weak aggressive. A sequence of analysis is done in this paper: data clean, data integration, data selection, data
transformation, Data Mining, and evolution is done
In this paper [11], driving features such as speed and acceleration distribution in the specific time intervals were
determined. For the input of data, Tachograph paper disk technique was used. A digital camera is used for
monitoring the exterior driving conditions. Next analysis was comparing disks of two drivers. The last analysis
is aimed to designate a change in speed profile resulting from driver fatigue. 1200 dpi scanning resolution and
disk with speed range 125km/hr allow obtaining accuracy about 0.004m/sec.
In this paper [12], the author describes a system to reduce accidents to a vast extends by checking eye blinking
of the driver which shows the drowsiness, obstacles situated in the road and the drunken condition of the
drivers. By utilizing the hardware platform comprising of Alcohol sensor MQ3, small scale controller, Li-Fi
framework, GSM module, ECU of vehicle. The planned framework would complete the capacity of speaking
with the base station by means of Li-Fi, GSM furthermore, control of different parameters. The entire Control
framework has the advantage of a little volume and high reliability.
In this [13] study, an accident avoidance strategy is outlined and a case study is examined considering road
accidents caused by drowsiness and sleepiness. Drowsiness or driver vigilance, in general, can be monitored
using a multi-sensor system to obtain estimates of the drowsiness level of the driver. The proposed design,
drowsiness checking framework is joined with a versatile and robust lateral controller. Drowsiness is viewed as
identified with the uncertainty in directing wheel directions for the vehicle lateral movement. Utilizing a robust
control theory plot, the uncertainties from road vehicle powers and driver inputs are tended to bringing about a
lateral controller. The controller is capable to re-shape the recurrence reaction of the vehicle in both side slip
angle increasing speed and side-slip point, moving the reaction into additional stable regions in the Nyquist

www.ijceronline.com Open Access Journal Page 36

 Intelligent Road Accidents Avoidance: A Review

diagram. An extra speed decrease concludes the total adjustment of the vehicle-driver framework. This
examination has two essential contributions to the current examine inactive safety framework development.
Initial, a driver observing framework is consolidated to a controller satisfying its ability for preventing the
accidents or moderating the effect. Second, the framework design may establish a response to the issue of the
expert progress between the controller and driver on account of imminent accidents.
In this paper [14], accurately predicting vehicle speed for an individual trip is a challenging topic because
vehicle speed is subjected to various factors such as route types, route curvature, driver behavior, and weather
and traffic condition. a big data based deep learning frame work is presented to address various factors such as
driver behavior, route type, traffic and weather condition, while predicting vehicle speed. This BDDL-SP
algorithm is applicable for trips with any route types. The input data for the trip of interest is extracted and
structured by big data management software, then this input information is used by a trained ANFIS model to
predict vehicle speed. This big data based deep learning speed prediction algorithm is implemented in
MATLAB .Experiment results indicate that the proposed algorithm is capable of accurately predicting vehicle
speed for both freeway and urban traffic networks. It take many data , it is not good for small data.
In this paper [15], the author introduced A fatigue assessment method based on the weighted Dempster-Shafer
(D-S) theory has been proposed by using the detected multi-source information such as the image, human pulse
and the steering wheel movement in an imperceptible way so as to enhance the fatigue monitoring accuracy and
system robustness. The data fusion results based on multiple sensor with enhanced fatigue monitoring accuracy
and system robustness have manifested more practical significance than that from single sensor in monitoring
driver fatigue.
In this paper [16], the author elaborate the risk level assessment based on the driving activities exceeding a
given threshold acceleration and speed value. This paper is organized as follows: Related work on sensor fusion
and Bayesian approach for prediction of risk is presented in Section II. The theory of belief functions is revisited
in Section III. In Section IV, we elaborate the procedure of risk level identification. Finally, some conclusions
are given. Statistical analysis and probability density function of three main risk levels such as Low, Medium
and High is derived using the vehicle collision property damage reports provided by a legal authority. A dataset
of 137 vehicles over 6 month’s period is studied. Evaluation results show that without significant reduction in
the total risk prediction accuracy, the ‘most-valuable’ group, High Risk level, prediction is achieved at an
acceptable level, and predictions’ accuracy of possible risk classes are distributed evenly.
This Paper [17] , provides a drowsiness detection system and a method that detects the driver's drowsiness in
real-time. Existing approaches have used vehicle-based and psychological measurements to detect the
drowsiness of the driver. In contrast to the beforehand determined issues, author proposed a system that
implements a non-intrusive technique for determining the driver’s fatigue. This system consists of a Raspberry-
Pi and a Pi camera module that continuously keeps scanning for facial landmarks. These landmarks are localized
using facial landmark detector and then the eye landmarks are used to calculate the eye aspect ratio (EAR). If
the EAR value decreases from the threshold value and the eyes remain closed for too long then the system
immediately alerts the driver with the aid of a buzzer. Furthermore, to ensure that the problem has been taken
care of, a notification is sent to the owner of the vehicle through e-mail when the driver dozes off for more than
a couple of times.
In this paper [18] the author proposed a system to prevent the car accident, the assistance strategy for safe
driving which focuses the normality of driver’s behavior. The driver’s behavior by using the steering wheel
operation has been examined. To validate the proposed method, fundamental experiments using Driving
Simulator system have been conducted. The experimental results of grabbing style for steering wheel and
statistical analysis using simple correlation between steering operation and vehicle velocity show the possibility
to estimate the internal state of drivers. And finally, Prototype of Steering wheel Sensing system has been
proposed based on these basic experiments and analysis.

III. COMPARATIVE ANALYSIS

Table1 depicts a comparative analysis of the several research papers that were reviewed as part of this
study.

Table1: Comparative Analysis of State-of-the-Art

S. No. Reference No. Parameters used Method Applied
1 4 Driving style GPS, accelerometer and smart phone is used. SVM
algorithm is used for pattern
2 5 Drowsiness, Pattern EMT is used
3 6 Driver’s behavior Using the steering wheel operation and driving
stimulator is used
4 7 Driver’s behavior Data mining and cognitive test

5 8 Drowsiness Using hough transform and Canny edge detection

www.ijceronline.com Open Access Journal Page 37

 Intelligent Road Accidents Avoidance: A Review

6 9 Driver’s behavior Using Gaze pattern , CNN and LSTM algorithms

are used
7 10 Driver’s behavior Cloud computing , decision tree is used

8 11 Speed Tachograph technology and digital camera are used

9 12 Drowsiness Alcohol sensor MQ3, Li-Fi framework , GSM

module and ECU of vehicle
10 13 Drowsiness Multi sensor system
11 14 Drowsiness BDDL-SP algorithm and ANFIS model
12 15 Fatigue and wheel turning Dempster-Shafer (D-S)
13 16 Fatigue Dampster shafer theory
14 17 Drowsiness Raspberry-Pi and a Pi
15 18 Driver’s behavior Driving Simulator system

IV. CONCLUSION
From the papers that were reviewed as part of this study it came to light that most of the methods
employ single parameter to detect the likelihood of road accidents. The driver’s behavior, fatigue, drowsiness
and speed were some of the commonly applied parameters. The results was also not accurate, whereas if we
introduce a system that will monitor drivers behavior using more than one parameter involving intelligent
machine learning algorithms we can have better accuracy in predicting the likelihood of occurrence of road
accident.

REFERENCES
[1]. https://www.who.int/news-room/fact-sheets/detail/road-traffic-injuries
[2]. https://commerce.gov.in/writereaddata/uploadedfile/MOC_636281140249481285_annual_report_16_17_eng.pdf
[3]. https://sites.ndtv.com/roadsafety/important-feature-to-you-in-your-car-5/
[4]. Bose, B., Dutta, J., Ghosh, S., Pramanick, P., & Roy, S. (2018, February). D&RSense: Detection of Driving Patterns and Road
Anomalies. In 2018 3rd International Conference On Internet of Things: Smart Innovation and Usages (IoT-SIU)(pp. 1-7). IEEE.
[5]. Golz, M., & Sommer, D. (2010, August). Monitoring of drowsiness and microsleep. In 2010 Annual International Conference of the
IEEE Engineering in Medicine and Biology(pp. 1787-1787). IEEE.
[6]. Imamura, T., Yamashita, H., Zhang, Z., bin OTHMAN, M. R., & Miyake, T. (2008, October). A study of classification for driver
conditions using driving behaviors. In 2008 IEEE International Conference on Systems, Man and Cybernetics (pp. 1506-1511).
IEEE.
[7]. Chakraborty, B., & Nakano, K. (2016, October). Automatic detection of the driver's awareness with the cognitive task from driving
behavior. In 2016 IEEE International Conference on Systems, Man, and Cybernetics (SMC) (pp. 003630-003633). IEEE
[8]. Katyal, Y., Alur, S., & Dwivedi, S. (2014, May). Safe driving by detecting lane discipline and driver drowsiness. In 2014 IEEE
International Conference on Advanced Communications, Control and Computing Technologies (pp. 1008-1012). IEEE.
[9]. Fan, X., Wang, F., Lu, Y., Song, D., & Liu, J. (2018, July). Eye Gazing Enabled Driving Behavior Monitoring and Prediction.
In 2018 IEEE International Conference on Multimedia & Expo Workshops (ICMEW) (pp. 1-4). IEEE.
[10]. Hwang, C. P., Chen, M. S., Shih, C. M., Chen, H. Y., & Liu, W. K. (2018, May). Apply Scikit-Learn in Python to Analyze Driver
Behavior Based on OBD Data. In 2018 32nd International Conference on Advanced Information Networking and Applications
Workshops (WAINA) (pp. 636-639). IEEE.
[11]. Rygula, A. (2009, June). Driving style identification method based on speed graph analysis. In 2009 International Conference on
Biometrics and Kansei Engineering (pp. 76-79). IEEE
[12]. Das, A., Ray, A., Ghosh, A., Bhattacharyya, S., Mukherjee, D., & Rana, T. K. (2017, August). Vehicle accident prevents cum
location monitoring system. In 2017 8th Annual Industrial Automation and Electromechanical Engineering Conference
(SEMICON) (pp. 101-105). IEEE.
[13]. .Boyraz, P., & Hansen, J. H. (2008, September). Active accident avoidance case study: integrating
[14]. drowsiness monitoring system with lateral control and speed regulation in passenger vehicles. In 2008
[15]. IEEE International Conference on Vehicular Electronics and Safety (pp. 293-298). IEE
[16]. Cheng, Z., Chow, M. Y., Jung, D., & Jeon, J. (2017, June). A big data based deep learning approach for vehicle speed prediction.
In 2017 IEEE 26th International Symposium on Industrial Electronics (ISIE) (pp. 389-394). IEEE.
[17]. Sanpeng, D., Xiaoli, X., Xuecui, Y., & Dehua, M. (2010, May). Research on the driver fatigue monitoring method based on the
Dempster-Shafer theory. In 2010 Chinese Control and Decision Conference (pp. 4176-4179). IEEE.
[18]. Gündüz, G., Yaman, Ç., Peker, A. U., & Acarman, T. (2017, June). Driving pattern fusion using dempster-shafer theory for fuzzy
driving risk level assessment. In 2017 IEEE Intelligent Vehicles Symposium (IV) (pp. 595-599). IEEE.
[19]. Hossain, M. Y., & George, F. P. (2018, October). IOT Based Real-Time Drowsy Driving Detection System for the Prevention of
Road Accidents. In 2018 International Conference on Intelligent Informatics and Biomedical Sciences (ICIIBMS) (Vol. 3, pp. 190-
195). IEEE.
[20]. Imamura, T., Yamashita, H., Zhang, Z., bin OTHMAN, M. R., & Miyake, T. (2008, October). A study of classification for driver
conditions using driving behaviors. In 2008 IEEE International Conference on Systems, Man and Cybernetics (pp. 1506-1511).
IEEE.

Noorishta Hashmi" Intelligent Road Accidents Avoidance: A Review" International Journal of

Computational Engineering Research (IJCER), vol. 09, no. 6, 2019, pp 35-38

www.ijceronline.com Open Access Journal Page 38