Head motion coefficient-based algorithm for distracted driving detection
| Pages | 171-188 |
| Published date | 01 April 2019 |
| DOI | https://doi.org/10.1108/DTA-09-2018-0086 |
| Date | 01 April 2019 |
| Author | Kwok Tai Chui,Wadee Alhalabi,Ryan Wen Liu |
| Subject Matter | Library & information science |
Head motion coefficient-based
algorithm for distracted
driving detection
Kwok Tai Chui
Department of Electronic Engineering,
College of Science and Engineering, City University of Hong Kong,
Kowloon, Hong Kong
Wadee Alhalabi
Department of Computer Science,
King Abdulaziz University, Jeddah, Saudi Arabia and
Virtual Reality Research Center, Effat University, Jeddah, Saudi Arabia, and
Ryan Wen Liu
School of Navigation, Wuhan University of Technology, Wuhan, China
Abstract
Purpose –Concentrationis the key to safer driving. Ideally,drivers should focus mainly on frontviews and
side mirrors.Typical distractions are eating,drinking, cell phone use, usingand searching things in car as well
as lookingat something outsidethe car. In this paper, distracteddriving detectionalgorithm is targeting onnine
scenariosnodding, head shaking,moving the head 45° to upper left and back to position,moving the head 45° to
lower leftand back to position, moving the head45° to upper right and back to position,moving the head 45° to
lower right and back to position, moving the head upward and back to position, head dropping down and
blinking as fundamental elements for distracted events. Thepurpose of this paper is preliminary study these
scenarios for the ideal distraction detection, the exact type of distraction.
Design/methodology/approach –The systemconsists of distraction detectionmodule that processes video
stream and compute motion coefficient to reinforce identification of distraction conditions of drivers. Motion
coefficient of thevideo frames is computed which follows bythe spike detection via statistical filtering.
Findings –The accuracy of head motion analyzer is given as 98.6 percent. With such satisfactory result, it is
concluded that the distraction detection using light computation power algorithm is an appropriate direction
and further work could be devoted on more scenarios as well as background light intensity and resolution of
video frames.
Originality/value –The system aimed at detecting the distraction of the public transport driver. By
providing instant response and timely warning, it can lower the road traffic accidents and casualties due to
poor physical conditions. A low latency and lightweight head motion detector has been developed for online
driver awareness monitoring.
Keywords Correlation analysis, Public transport, Social good, Road safety, Distracted driving,
Head motion
Paper type Research paper
1. Introduction
Nowadays, people are relying heavily on the public transport system as well as self-driving.
According to the World Health Organization, nearly 1.3m people were killed and 50m people
were injured in road traffic collision (World Health Organization, 2013), which cause heavy
burden to the health system. It is estimated that 3 percent of the respective GDP of the world
(more than 3,000bn USD) is lost due to traffic crashes (World Health Organization, 2009).
Furthermore, accidents also double the car insurance cost. Poor physical conditions like
drowsiness, stress, distraction, drink, drug, sudden illness, mental defect and
overrepresentation of drivers are typical reasons of serious traffic accidents (Chen and
Donmez, 2016; Papantoniou et al., 2017). Generally, the existing works devote major efforts
on distracted detection, stress detection and drowsiness detection in which distracted
Data Technologies and
Applications
Vol. 53 No. 2, 2019
pp. 171-188
© Emerald PublishingLimited
2514-9288
DOI 10.1108/DTA-09-2018-0086
Received 30 September 2018
Revised 26 January 2019
Accepted 23 February 2019
The current issue and full text archive of this journal is available on Emerald Insight at:
www.emeraldinsight.com/2514-9288.htm
171
Head motion
coefficient-
based
algorithm
detection is the focus of this paper. Readers who are interested in driver stress detection and
drowsiness detection are suggested to refer to (Affanni et al., 2018; Chen et al., 2017) and
(Awais et al., 2017; Chui et al., 2016; McDonald et al., 2018), respectively.
More than 30 percent of driver had reported distracted driving (Schroeder et al., 2018) in
2010–2015. Although some measures like law banning texting (Kim, 2018) and talking on
mobile phone while driving (Oviedo-Trespalacios et al., 2017), safe driving slogan and
educational messages have been proposed to lower the chances of distracted driving, the
effectiveness has remained low. The public transport system is indispensable for most
people. As road users (pedestrian, cyclist or motorist), safe publication transport is the major
concern of social good (Rathore et al., 2017). Prevention is always better than cure to achieve
social inclusive economic growth for sustainability (Visvizi et al., 2018). A distracted
detection system via head motion coefficient-based algorithm has been proposed that keeps
track on the driver’s vigilance to prevent road traffic collision. It is suggested to initialize the
auto-navigator of the vehicle (if available) which gives a suitable advice to the driver for
some critical cases. It is envisaged that the diminishing accidents and casualties will yield an
inexpensive third party motor insurance cost that benefits public transport companies,
drivers and passengers. A low latency head motion detector has been developed for online
driver awareness monitoring.
The rest of paper is organized as follows: Section 2 summarizes the related works of
distracted driving detection. The system overview is presented in Section 3 and the
methodology of driver distraction detection is explained and formulated in Section 4. Then,
performance evaluation of the proposed algorithm is given in Section 5. Finally, a conclusion
is drawn in Section 6.
2. Related works
Macroscopically, driving style is influenced by two factors named environmental factors
and human factors (Martinez et al., 2018). Examples of former are traffic, weather, other
drivers and road condition. The latter includes age, gender, driving experience, fatigue,
stress and distraction. In the following, driver distraction is the main focus.
Driver distraction detection can be categorized into four types: visual; cognitive;
auditory and biomechanical distraction National Highway Traffic Safety Administration
by Ranney et al. (2001). In this paper, authors would like to focus on the first type of
distraction. A review article is recommended for readers who are interested in the wide
range of detail for drive inattention monitoring that covers both driver distraction and
fatigue (Dong et al., 2011).
Facial feature tracking, head pose and gaze estimation and 3-D geometric reasoning were
adopted to detect the eyes off the road (Vicente et al., 2015). It is noted that this method not
only detect the head pose but also the eyebrows, eyes and mouth of the driver. The accuracy
was over 90 percent in 18 targeted locations/scenarios.
A support vector machine (SVM) was employed for binary class driver distraction
detection (Liang et al., 2007). Generally, the algorithm detected the eye movement by
analyzing the sequences of smooth pursuits, saccades and fixation using the chacteristics of
dispersion and velocity. Ten volunteers were seclected for the evaluation and the algorithm
yielding 81 percent of average accuracy (without cross validation). It is noted that the
distracted events were not mentioned.
Another work (Tawari et al., 2014) was employing three techniques for driver attention
detection. Continuous head movement estimator, field of view estimation for attention and
attention buffer modulation for head pose monitoring and distraction duration estimation.
Results and performance evaluation were not provided.
Futher performance comparison between proposed work and realted works will be
discussed in Section 5.4.
172
DTA
53,2
To continue reading
Request your trial